DOI10.1007/s10479-009-0603-7
Re-solving stochastic programming models for airline revenue management
Lijian Chen·Tito Homem-de-Mello
?Springer Science+Business Media,LLC2009
Abstract We study some mathematical programming formulations for the origin-destin-ation model in airline revenue management.In particular,we focus on the traditional prob-abilistic model proposed in the literature.The approach we study consists of solving a se-quence of two-stage stochastic programs with simple recourse,which can be viewed as an approximation to a multi-stage stochastic programming formulation to the seat alloca-tion problem.Our theoretical results show that the proposed approximation is robust,in the sense that solving more successive two-stage programs can never worsen the expected rev-enue obtained with the corresponding allocation policy.Although intuitive,such a property is known not to hold for the traditional deterministic linear programming model found in the literature.We also show that this property does not hold for some bid-price policies. In addition,we propose a heuristic method to choose the re-solving points,rather than re-solving at equally-spaced times as customary.Numerical results are presented to illustrate the effectiveness of the proposed approach.
Keywords Stochastic programming·Multi-stage models·Revenue management
1Introduction
Revenue management involves the application of quantitative techniques to improve pro?ts by controlling the prices and availabilities of various products that are produced with scarce resources.Perhaps the best known revenue management application occurs in the airline Supported by the National Science Foundation under grant DMI-0115385.
L.Chen
Department of Industrial Engineering,J.B.Speed School of Engineering,University of Louisville, Louisville,KY40292,USA
e-mail:lijian.chen@https://www.doczj.com/doc/633246601.html,
T.Homem-de-Mello( )
Department of Mechanical and Industrial Engineering,University of Illinois at Chicago,Chicago,
IL60607,USA
e-mail:thmello@https://www.doczj.com/doc/633246601.html,
industry,where the products are tickets(for itineraries)and the resources are seats on?ights. In view of many successful applications of revenue management in different areas,this topic has received considerable attention in the past few years both from practitioners and academics.The recent book by Talluri and van Ryzin(2004)provides a comprehensive introduction to this?eld,see also references therein.
A common way to model the airline booking process is as a sequential decision problem over a?xed time period,in which one decides whether each request for a ticket should be accepted or rejected.A typical assumption is that one can separate demand for individual itinerary-class pairs;that is,each request is for a particular class on a particular itinerary, and yields a pre-speci?ed fare.Typically,a class is determined by particular constraints associated with the ticket rather than the physical seat.For example,a certain class may require a14-day advance purchase,or a Saturday night stay,etc.(for the purposes of this paper,we assume that?rst-class seats are allocated separately).
The existence of different classes re?ects different customer behaviors.The classical example is that of customers traveling for leisure and those traveling on business.The former group typically books in advance and is more price-sensitive,whereas the latter behaves in the opposite way.Airline companies attempt to sell as many seats as possible to high-fare paying customers and at the same time avoid the potential loss resulting from unsold seats. In most cases,rejecting an early(and lower-fare)request saves the seat for a later(and higher-fare)booking,but at the same time that creates the risk of?ying with empty seats. On the other hand,accepting early requests raises the percentage of occupation but creates the risk of rejecting a future high-fare request because of the constraints on capacity.
Many of early models were built for single?ights.While that environment allows for the derivation of optimal policies via dynamic programming even with the incorporation of extra features(Subramanian et al.1999),the drawback is clear in that the booking policy is only locally optimized and it cannot guarantee global optimality.Because of that,models that can incorporate a network of?ights are usually preferred by the https://www.doczj.com/doc/633246601.html,work models, however,can only provide heuristics for the booking process,since determining the optimal action for each request in a network environment is impractical from a computational point of view.One type of heuristics is based on mathematical programs,where the decision variables are the number of seats to allocate to each class.In particular,methods based on linear programming techniques have been very popular in industry,for several reasons:?rst, linear programming is a well developed method in operations research;its properties have been thoroughly studied for decades.Secondly,commercial software packages for linear programming are widely available and have proven ef?cient and reliable in practice.Finally, the dual information obtained from the linear program can be used to derive alternative booking policies,based on bid prices;we will return to that in Sect.4.
Booking methods based on linear programming were thoroughly investigated by Williamson(1992).The basic models take stochastic demand into account only through expected values,thus yielding a deterministic program that can be easily solved.However, the drawback of such approach is obvious,as it ignores any distributional information about the demand.A common way to attempt to overcome that problem is to re-solve the LP sev-eral times during the booking horizon.While such an approach may seem intuitive,it turns out that re-solving can actually back?re—indeed,Cooper(2002)shows a counter-example where re-solving the LP model may lower the total expected revenue.
An alternative way to incorporate demand distribution information into the model is by formulating a stochastic linear program.In the particular case of airline bookings,such models typically reduce to simple recourse models,a formulation that is called probabilistic nonlinear program in the revenue management literature(see,e.g.,Williamson1992).Higle
and Sen(2006)propose a different stochastic programming model,based on leg-based seat allocations,which yields an alternative way to compute bid prices.Another stochastic opti-mization model is proposed by Cooper and Homem-de-Mello(2007),who consider a hybrid method where the second stage is actually the optimal value function of a dynamic program.
In this paper we discuss various aspects of the multi-stage version of the simple recourse model discussed above(henceforth denoted respectively MSSP and SLP).The MSSP model we present is shown to yield a better policy than SLP in terms of expected total revenue under the corresponding allocation policies;however,that multi-stage model does not have convexity properties(even its continuous relaxation),whereas simple recourse models can be solved very ef?ciently with linear integer programming.Of course,these conclusions are valid for the underlying MSSP model;alternative multi-stage models proposed in the literature(notably the ones in DeMiguel and Mishra(2006)and M?ller et al.(2004))do not suffer from the non-concavity issue,although a precise relationship with the SLP model is not established in those papers.
Given the dif?culty to develop exact algorithms for large multi-stage problems,we pro-pose an approximation based on solving a sequence of two-stage simple recourse models. The main advantage of such an approach is that,as mentioned above,each two-stage prob-lem can be solved very ef?ciently,so an approximating solution to the MSSP can be ob-tained reasonably quickly.The idea of solving two-stage problems sequentially is not new, and appears in the literature under names such as rolling horizon and rolling forward;see, for instance,Balasubramanian and Grossmann(2004),Bertocchi et al.(2006),Kusy and Ziemba(1986).The details on the implementation of the rolling horizon,however,vary in the above works.Our work is more closely related to Balasubramanian and Grossmann (2004)in that we consider shrinking horizons,i.e.,each two-stage problem is solved over a period spanning from the current time until the end of the booking horizon.In this paper this is called the re-solving SLP approach.
Although the rolling horizon approach has been proposed in the literature,to the best of our knowledge there have been no analytical results regarding the quality of the approxima-tion.In this paper we provide some results of that nature,though we do not claim to give de?nitive answers.More speci?cally,we compare the policies obtained from the re-solving SLP approach with the policies from the MSSP model.We show that,for a given partition into stages,the policy from MSSP is better than the policy from re-solving SLP.However, the inclusion of just one extra re-solving point can make the re-solving approach better.The importance of this conclusion arises from the fact that,because of the sequential nature of the re-solving procedure,adding an extra re-solving point requires little extra computational effort;in comparison,including an extra stage in a multi-stage model makes the problem considerably bigger and therefore harder to solve.
We also study the effect of re-solving the SLP model,compared to not re-solving it. Our results show that,unlike the aforementioned example in Cooper(2002)for the DLP model,solving the SLP sequentially cannot be worse(in terms of expected revenue from the resulting policy)than solving it only once.In addition,we provide an example to illustrate that re-solving may actually be worse in the context of standard bid-price policies,where the bid prices are calculated from the dual problem of either the DLP or the SLP models. These results are,to the best of our knowledge,novel.
Motivated by the?exibility of the re-solving approach,we also study the issue of whether one can improve the results by carefully choosing the re-solving points instead of using equally-sized intervals as it is usually done.Indeed,the structure of our problem allows us to do so,and we provide a heuristic algorithm to determine the re-solving points.Our numerical results,run for two relatively small-sized networks,indicate that the procedure is effective.
The remainder of the paper is organized as follows.In Sect.2we introduce the notation and describe mathematical programming methods for the seat allocation problem.The re-solving approach is treated in detail in Sect.3,and bid-price policies are discussed in Sect.4. Section5describes the algorithm for improving the choice of re-solving points,whereas Sect.6presents numerical results.Concluding remarks are presented in Sect.7.
2Allocation methods
Following the standard models in the literature,we consider a network of?ights involving p booking classes of customers.This model can represent demand for a network of?ights that depart within a particular day.Each customer requests one out of n possible itineraries,so we have r:=np itinerary-fare class combinations.The booking process is realized over a time horizon of lengthτ.Let{N jk(t)}denote the point process generated by the arrivals of class-k customers who request itinerary j.Typical cases customary in the revenue management literature are(i){N jk(t)}is a(possibly nonhomogeneous)Poisson process,and(ii)there is at most one unit of demand per time period.Throughout this paper we do not make those assumptions unless when explicitly said otherwise;we only assume that arrival processes corresponding to different pairs(j,k)are independent of each other.
The demand for itinerary-class(j,k)over the whole horizon is denoted byξjk(i.e.,ξjk= N jk(τ))and we denote byξthe whole vector(ξjk).The network is comprised of m leg-cabin combinations,with capacities c:=(c1,...,c m),and is represented by an(m×np)-matrix A≡(a i,jk).The entry a i,jk∈{0,1}indicates whether class k customers use leg i in itinerary j.
Most policies we deal with in this paper are of allocation type.We denote by x jk the de-cision variable corresponding to the number of seats to be allocated to class k in itinerary j. Whenever an itinerary-class pair(j,k)is accepted,the revenue corresponding to the fare f jk accrues.A customer’s request is rejected if no seats are available for his itinerary-class, in which case no revenue is realized.The vectors of decision variables and fares are denoted respectively by x=(x jk)and f=(f jk).
Allocation methods require solving an optimization problem to?nd initial allocations before the booking process starts.The classical deterministic linear program(DLP)for the network problem is written as follows:
max{f T x:Ax≤c,x≤E[ξ],x≥0}.[DLP] Model[DLP]is well known in the revenue management literature.Implementation of the
resulting policy x?is very simple—accept at most x?
jk class k customers in itinerary j.
Notice that the policy is well-de?ned even if the solution x?is not integer.Notice also that the objective function of[DLP]is not the actual revenue resulting from the policy x?,even if x?is integer.
A major drawback of formulation[DLP]is the fact that it completely ignores any infor-mation regarding the distribution of demand,except for the mean.This leads to the stochastic programming formulation
max{f T E[min{x,ξ}]:Ax≤c,x∈Z+},
where the min operation is interpreted componentwise(note that we have imposed an inte-grality constraint on x to ensure we get an integer allocation of seats;we will comment on
that later).Equivalently,we have
max f T x+E[Q(x,ξ)]
[SLP]
subject to:Ax≤c,x∈Z+,
where Q(x,ξ)=max{?f T y|x?y≤ξ,y≥0}.Notice that[SLP]can be formulated as a two-stage integer problem with simple recourse.A major advantage of such models is that,whenξhas a discrete distribution with?nitely many scenarios,problem[SLP]can be easily solved because of its special structure.In principle,this may not be the case of our model,for example when the total demand for each itinerary-class pair(j,k)has Poisson distribution,which has in?nite support.It is clear,however,that in that case all but a?nite number of points in the distribution have negligible probability;thus,we can approximate the distribution ofξjk by a truncated Poisson.Thus,in what follows we assume thatξtakes on?nitely many values,and that those values are integer.
To describe the solution procedure,we need to introduce some notation.For each itinerary-class pair(j,k),let S jk denote the number of possible values taken byξjk,and let
d1 jk ,...,d S jk
jk
denote those values,ordered from lowest to highest.Let{δs
jk
},s=1,...,S jk
be coef?cients de?ned asδs
jk :=f jk P(ξjk≤d s jk).As discussed in Birge and Louveaux
(1997)and Kall and Wallace(1994),problem[SLP]can then be re-written as
max x,u,u0f T x?
j,k
S jk
s=1
δs
jk
u s
jk
subject to:Ax≤c,
u0 jk +
S jk
s=1
u s
jk
?x jk=?E[ξjk],
u0
jk
≤d1jk?E[ξjk],
0≤u≤1,
x∈Z+.
(1)
Notice that the decision variables of the linear integer program(1)are the vectors x=(x jk), u0:=(u0jk)and u:=(u s jk),s=1,...,S jk(the vectors u and u0correspond to the slopes of the objective function of the second stage,which is piecewise linear).That is,problem(1) has O(Snp)variables and O(Snp+m)constraints(where S:=max j,k S jk),which is far smaller than the deterministic linear program corresponding to general two-stage programs with?nitely many scenarios—in that case,it is well known that the number of constraints and variables is exponential on the number of scenarios,see for instance Birge and Louveaux (1997).Thus,problem(1)can be solved by standard linear integer programming software.
It is worthwhile pointing out that,if one implements the booking policy based on seat allocations(we call it the allocation policy henceforth),then the objective function of[SLP] does correspond to the actual expected revenue resulting from a feasible policy x—though this is not true if the integrality constraint is relaxed.Note also that the solution obtained from[DLP]yields the same expected revenue as its rounded-down version.Moreover,it is easy to see that a rounded-down feasible solution to[DLP]is feasible for[SLP].An immediate consequence of these facts is that the optimal allocation policy calculated from [SLP]is never worse than that of the DLP model in terms of expected total revenue.We emphasize that this is true in the present context of simple allocation policies,so such a conclusion may not hold for other policies.
2.1Multi-stage models
We discuss now a multi-stage version of the SLP model described above,in which the
policy is revised from time to time in order to take into account the information about de-
mand learned so far.Suppose we divide the time horizon[0,τ]into H+1stages numbered
0,1,...,H.The stages correspond to some partition0=t0 h(h=1,...,H)consists of time interval(t h?1,t h].The decision variables at each stage are denoted x0,...,x H,where x h=(x h jk).Also,we associate with each stage h,h=1,...,H, random variablesξh jk representing the total demand for itinerary-class(j,k)between stages h?1and h,that is,ξh jk =N jk(t h)?N jk(t h?1).We denote byξh the random vector(ξh jk). Notice that the decision vector x h at stage h is actually a function of x0,x1,...,x h?1and ξ1,...,ξh. The resulting multi-stage model is written as follows: max f T x0+Eξ1 Q1(x0,ξ1) [MSSP] subject to:Ax0≤c,x0∈Z+. The function Q1is de?ned recursively as Q h(x0,...,x h?1,ξ1,...,ξh) =max x h f T x h?f T[x h?1?ξh]++Eξh+1 Q h+1(x0,...,x h,ξ1,...,ξh+1) subject to Ax h≤c?A h m=1 min{x m?1,ξm},x h∈Z+ (h=1,...,H?1),where[a]+:=max{a,0}and the max and min operations are interpreted componentwise.Notice that we use the notation Eξh+1[Q h+1(x0,...,x h,ξ1,...,ξh+1)]to indicate the conditional expectation E[Q h+1(x0,...,x h,ξ1,...,ξh+1)|ξ1,...,ξh].For the ?nal stage H we have Q H(x0,...,x H?1,ξ1,...,ξH)=?f T x H, where x H=[x H?1?ξH]+. Observe that in the above formulations we use the equality E[min{a,Y}]=a?E[[a?Y]+], where a is deterministic and Y is a random variable. From the discussion above,we see that the multi-stage stochastic program[MSSP]is, in principle,a better model for the problem under study than[DLP]or[SLP],since it takes the stochastic and dynamic features of the problem into account.Several methods have been developed for multi-stage stochastic programs where all stages involve linear problems,see for instance Birge(1985),Birge et al.(1996),Gassmann(1990).One re-quirement for these algorithms to work is that the expected recourse function be convex (or concave,for a maximization problem).Moreover,these algorithms were devised for continuous problems—research on multi-stage integer problems is ongoing.Unfortunately, model[MSSP]described above does not?t that framework.In fact,as we show below, even the continuous relaxation of that problem does not have a concave expected recourse function. Example1Consider a three stage programming problem corresponding to a single-leg,two-class problem,with the capacity equal to4.There are two independent time periods.The demands for the two time periods are deterministic,respectively?ξ1=(1,2)and?ξ2=(2,2). The fares for each class are respectively f1=$500and f2=$100.It is easy to check that Q1((1,1),?ξ1)=1000and Q1((1,3),?ξ1)=400.However,the average of these values is 1400/2=700,which is larger than500,the value of Q1at the midpoint(1,2).Therefore, Q1is not concave. It is worth noticing that the lack of concavity exists only for problems with three or more stages,the reason for that being the min operation in the constraint of the problem de?ning Q1.Also,it is important to keep in mind that this issue arises only in the formulation we have presented.For example,DeMiguel and Mishra(2006)circumvent that problem by proposing a model which is“partially non-anticipative”,in the sense that the decision maker has perfect information about the current stage(but not about future ones);M?ller et al.(2004),on the other hand,use a different set of decision variables—cumulative allocations rather then per-stage ones—and obtain a multi-stage integer linear stochastic program,which without the integrality constraints yields concave functions(indeed,when the number of stages is two, the model in M?ller et al.(2004)is just[SLP]).Those models,however,can still be dif?cult to solve,especially for a large network with a large number of scenarios;to avoid that,in Sect.3we will discuss an alternative approach to solve the allocation problem. 3Re-solving SLP model A natural alternative to the multi-stage approach described in Sect.2.1is to revise the book-ing policy from time to time by re-solving a simpler model such as[SLP].While intuitively it can be argued that such an approach may yield policies that are inferior to the ones given by[MSSP]—which by construction?nds the optimal dynamic seat allocation policy—we believe that the re-solving approach is worth studying,for several reasons.First,as discussed before,the solution of[MSSP]is likely to be computationally intensive,especially for large networks.Re-solving a problem such as[SLP],on the other hand,is reasonably fast,since as described earlier each instance of[SLP]is equivalent to a moderately sized linear integer program.Moreover,it is clear that the complexity of[MSSP]increases rapidly as the num-ber of stages grow,since the number of decision variables and constraints becomes larger; the complexity of re-solving,on the other hand,clearly grows linearly with the number of stages. We formalize now the re-solving approach.As in the case of the multi-stage model,we partition the booking time horizon[0,τ]into segments{0},(0,t1],(t1,t2],...,(t H?1,τ]with 0=t0 We initially(i.e.at time0)solve the following problem: max f T E min x, H m=1 ξm [SLPR-0] subject to Ax≤c,x∈Z+. Let x 0denote the optimal solution of the above problem.Then,at each time t h ,h =1,...,H ?1we solve max f T E min x,H m =h +1 ξm ξ1=?ξ 1,...,ξh =?ξh [SLPR-h ]subject to Ax ≤c ?A h m =1min {x m ?1,?ξ m },x ∈Z +and denote the optimal solution by x h .Note that the above model makes use of the infor-mation up to time t h —this is the reason why the constraints involve the realizations {?ξm }instead of the random variables {ξm }.Notice also that the expectation in the objective func-tion of [SLPR-h ]is calculated with respect to the overall demand from period h +1on.The idea is that this would be the problem one would solve if it was assumed that no more re-solving would occur in the future. The idea of re-solving an optimization model to use the available information is not new.In the context of stochastic programming,this is sometimes called the rolling forward approach in the literature,see for instance Bertocchi et al.(2006)and Kusy and Ziemba (1986).The speci?c aspects of each problem,however,lead to different ways to implement the rolling mechanism.For example,in Bertocchi et al.(2006)a two-stage model is initially solved where the realizations at the second stage correspond to all possible scenarios of the original problem.That is,such a problem can be enormous and must be solved via some sampling method (see Shapiro 2003for a compilation of results).In contrast,in our case the two-stage program [SLPR-0]deals only with the total demand H m =1ξm .This reduces the number of scenarios drastically,especially when the distribution of H m =1ξm can be de- termined directly.Such is the case,for example,when the demand for each itinerary-class (j,k)arrives according to a Poisson process—in that case, H m =1ξm has Poisson distribu-tion,which then can be truncated as discussed in Sect.2.This results in tractable two-stage simple recourse problems that can be solved exactly. The approach of considering two-stage problems with increasingly small horizons has been used by Balasubramanian and Grossmann (2004)in the context of a multi-period scheduling problem.They call it the shrinking horizon framework .Their motivation is simi-lar to ours—to provide a practical scheme for a dif?cult multi-stage integer problem.Notice however that in our case we have the additional issue of lack of convexity,as discussed in Sect.2.1. The idea of re-solving an optimization model to take advantage of the information accu-mulated so far has been a common practice in revenue management applications.As men-tioned earlier,however,Cooper (2002)shows a simple counter example where re-solving [DLP]leads to a worse policy (in terms of expected revenue)than what would be obtained if one had kept the original policy throughout.As we show next,this does not happen in case [SLP]is re-solved. Proposition 1The allocation policy obtained from using models [SLPR-h ],h =0,...,H ?1,yields an expected revenue which is bigger or equal to that given by the allocation policy obtained from solving [SLPR-0]only . Proof It suf?ces to show that the ?rst re-solving cannot worsen the expected revenue.Let x 0denote the optimal solution of [SLPR-0].Then,during the time interval (0,t 1],the booking allocation policy is x0regardless of whether we are re-solving or not.The difference happens after time t1.For non-re-solving process,the policy to be used from time t+ 1 on is x1:= x0?min{x0,?ξ1}because the non-re-solving process continues to apply the initial policy. The policy for the re-solving process is obtained by solving(SLPR-1),i.e., max f T E min x, H m=2 ξm ξ1=?ξ1 subject to Ax≤c?A min{x0,?ξ1},x∈Z+. Notice that,by de?nition of x1, Ax1=Ax0?A min{x0,?ξ1}≤c?A min{x0,?ξ1} for any possible realization ofξ1.That is,x1is a feasible solution for the re-solving model, which means that the policy obtained by re-solving cannot be worse than x1.Since the objective function of[SLPR-1]is the expected revenue from time t+ 1on,it follows that re-solving cannot yield worse results than the initial policy. Proposition1shows that,in terms of expected revenue under the corresponding allo-cation policies,solving[SLPR-h]successively will keep improving(though perhaps not strictly)the booking policy.Notice that the re-solving method is somewhat similar to the multi-stage model in the sense that both yield dynamic booking policies that take the in-formation available so far into account.As mentioned earlier,it is intuitive that in general [MSSP]gives a better solution.The proposition below formalizes that result. Proposition2Under the same partition setting,the allocation policy from multi-stage model[MSSP]yields an expected revenue which is bigger or equal to that given by the allocation policy obtained from solving[SLPR-h]successively. Proof Let h be the set of all possible sample paths of(ξ1,...,ξh).A feasible solution for problem[MSSP]has the form x0× H h=1 ?ξ1,...,?ξh∈ h x h(?ξ1,...,?ξh),(2) where indicates the Cartesian product and each component x h(?ξ1,...,?ξh)satis?es Ax h(?ξ1,...,?ξh)≤c?A h m=1 min{x m?1(?ξ1,...,?ξm?1),?ξm}.(3) On the other hand,consider model[SLPR-h]under a speci?c realizationˉξ1,...,ˉξh in the scenario tree,and denote its optimal solution byˉx h(ˉξ1,...,ˉξh).Consider the vector formed by the Cartesian product of such solutions for all realizations and all stages,i.e., ˉx0× H h=1 ˉξ1,...,ˉξh∈ h ˉx h(ˉξ1,...,ˉξh). It is clear that the resulting vector has the form (2).Moreover,since the [SLPR-h ]problem has the constraints Ax h ≤c ?A h m =1min {x m ?1,ˉξm },it follows that ˉx h (ˉξ1,...,ˉξh )satis-?es (3).Therefore,the combined solution from the [SLPR-h ]models is actually a feasible solution in [MSSP]. It is important to notice that Proposition 2is valid under the same partition into stages.Indeed,the ?exibility of the re-solving approach allows for the inclusion of additional re-solving points without much burden—in other words,the complexity grows linearly with the number of stages,which in general is not true for the multi-stage model.It is natural then to compare the MSSP model and the re-solving SLP model with a re?ned partition.As the example below shows,including even one extra re-solving point may yield better results than solving the multi-stage model. Example 2Consider a single-leg problem with two independent booking classes,1and 2,with f 1=$300,f 2=$200.The capacity is equal to 15,and the booking time horizon has three time periods,1,2,3.During period 1,the distribution of demand for classes 1and 2is ξ11= 0with probability 1 2,1with probability 12 ,ξ12=0with probability one.Likewise,the distribution of demand during period 2is ξ21= 3with probability 12,7with probability 12 ,ξ22= 3with probability 12,5with probability 12,and ξ31= 5with probability 12 ,7with probability 12,ξ32= 4with probability 12 ,8 with probability 12,during period 3. Because of the limited scale of this problem,the multi-stage model can be solved by enu-meration.Suppose we solve a three-stage problem with the second and third stages de?ned respectively as time intervals (0,1]and (1,3].It is easy to check that the optimal solution from this model is x 0=(15,0)T ,x 11=(10,5)T (when ξ11=0happens)and x 11=(10,4)T (when ξ11=1happens).The expected total revenue is $3900.For the re-solving SLP approach,x 0=(9,6)T is the ?rst stage decision.Although this solution does not coincide with that from the MSSP model,it turns out that,once we re-solve at time 1,we obtain the same expected revenue of $3900resulting from [MSSP].When we include an extra re-solving point at time 2,the expected total revenue becomes $4000,which is $100,or 2.56%higher than that of MSSP model.For a large network,the improvement would be signi?cant.This example suggests that applying the re-solving procedure can be more bene?cial (in terms of expected revenue under the allocation policy)than solving a more complicated multi-stage model. Another type of comparison between the MSSP model and its re-solving counterpart can be made when one has perfect information.The proposition and corollary below show that,in that case,re-solving is in fact optimal.Although the direct applicability of these results is limited (as the true problem is stochastic),the proofs illustrate the relationship between the two approaches. Proposition 3Under perfect information ,the policies given by models [SLPR-0]and [MSSP]are equivalent ,in the sense that they yield the same expected revenue . Proof Suppose there is perfect information,i.e.there is only one sample path,which we denote by (ˉξ1,...,ˉξH ).Each decision x h is made with knowledge of the whole vector (ˉξ 1,...,ˉξH ).Then,[MSSP]is written as max H h =1f T min {x h ?1,ˉξh }subject to Ax 0≤c,(4) Ax 1≤c ?A min {x 0,ˉξ1},(5) Ax 2≤c ?A min {x 0,ˉξ 1}?A min {x 1,ˉξ2},(6)... Ax H ?1≤c ?A H ?1 h =1min {x h ?1,ˉξh },(7) x ∈Z +. It is clear from the above formulation that,if ?x :=(?x 0,...,?x H ?1)is an optimal solution for (4),then so is (min {?x 0,ˉξ 1},...,min {?x H ?1,ˉξH }).Since the region de?ned by each con-straint (5)–(7)contains the region de?ned by the next inequality (recall that A has only non-negative entries),it follows that the above problem can be simpli?ed to max f T H h =1x h ?1subject to A H h =1 x h ?1≤c,x h ?1≤ˉξ h ,h =1,...,H,x ∈Z +. (8)Consider now the problem max f T y subject to Ay ≤c,y ≤H h =1ˉξh ,y ∈Z +. (9) Notice that (9)is precisely problem [SLPR-0]under perfect information.Thus,to show the property stated in the proposition,it suf?ces to prove that the policies derived from (8)and (9)are equivalent.To do so,de?ne C y :={x feasible in (8): H h =1x h ?1=y }.Let F be a mapping from {C y :y ∈R np }into the feasible region of (9),de?ned as F (C y )=y .Consider now the mapping G that represents the application of the policy obtained from [SLPR-0].We can express G as a mapping from the feasible region of (9)into R H ×np ,de?ned as follows.For each pair(j,k),let a jk≤H be the largest number such that y jk≥ a jk h=1 ˉξh jk . Then,we de?ne G as G h?1(y)jk:=? ?? ?? ˉξh jk ,1≤h≤a jk, y jk? h?1 m=1 ˉξm jk ,h=a jk+1, 0,a jk+2≤h≤H. Notice that from the above de?nition we have0≤G h?1(y)jk≤ˉξh jk for all h=1,...,H. Moreover, H h=1G h?1(y)jk= a jk h=1 ˉξh jk +y jk? a jk m=1 ˉξm jk + H h=a jk+2 0=y jk. It follows that G(y):=(G0(y),...,G H?1(y))is feasible for(8)and,in addition, H h=1G h?1(y)=y.That is,G(y)∈C y.By viewing C y as an equivalence class,we see that G is actually the inverse of F,i.e.,F is a one-to-one mapping.Now,since F preserves the objective function value,we conclude that the policies from(8)and(9)are equivalent. Propositions1,2and3together yield the following result. Corollary1Under the same partition into stages and perfect information,the policies given by models[SLPR-h],h=0,...,H?1and[MSSP]are equivalent. 4Bid-price methods The policies discussed in the previous sections are of allocation type—i.e.,accept customers from a certain class until the corresponding allocations are used up.Another well-known type of policies is given by bid prices.In the context of airline booking,this means each leg has an incremental price.A booking request corresponds to seat occupation in one or more legs;the sum of the incremental prices for those legs is called bid price for this request. Then,the request is accepted only if its fare is bigger or equal to that amount.Notice that this method automatically provides a form of“nesting”even in a network environment,since by construction it cannot happen that a low-fare customer is accepted while a high-fare request is rejected. A common way to determine bid prices is through the dual variables of the allocation problems discussed in Sect.2.Williamson(1992)studies the case where the bid prices are the dual variables of[DLP].This method is quite simple and easy to use.However,as pointed out in Talluri and van Ryzin(1999),it may behave poorly.A natural alternative is to look at the dual multipliers of[SLP].A third method,proposed by Talluri and van Ryzin (1999),calculates the dual multipliers of[SLP]under perfect information and averages the resulting values over a number of samples.Another approach is described in Higle and Sen (2006),where the dual multipliers are calculated from a different stochastic programming problem(a leg-based seat allocation formulation).We refer to Talluri and van Ryzin(2004) for other approaches to compute bid prices. It is clear from the structure of the bid-price policy that its form is too rigid—depending on the values of the bid prices,entire classes may be rejected.In practice,the bid prices are re-calculated on a regular basis in order to take into account new information about the demand,thus providing a more?exible policy.When the bid prices are obtained from dual multipliers of a mathematical program,this amounts to re-solving the problem with updated information,which is precisely the setting of Sect.3. In light of the results of Sect.3,a natural question that arises is whether the expected revenue under a bid price policy can be guaranteed not to worsen with a re-solving approach. Unfortunately,the answer is negative,even if the bid prices are calculated from the SLP model.We show below a small example to illustrate this issue. Example3Consider a one-leg model with two booking classes,the less price-sensitive customers(class1)paying$100and more price-sensitive customers(class2)paying$60. The capacity is4.The demands for those customers are denoted byξ1,ξ2.Therefore the DLP model is max{100x1+60x2:x1+x2≤4,0≤x i≤E[ξi]}. For an arbitrary time t≤τ,letξt k denote the(random)number of arrivals of class k up to time t,and let?ηt denote the actual number of sold seats up to time t.Therefore,the re-solving DLP model is max{100x1+60x2:x1+x2≤4??ηt,0≤x i≤E[ξi?ξt i]}. The tables below show the possible outcomes of the corresponding bid price policy,ac-cording to the values of the various quantities involved in the above problems. Case1E[ξ1]>4and E[ξ1?ξt1]>4??ηt: Booking methods Acceptable classes through time t Acceptable classes thereafter DLP without re-solving11 DLP with re-solving11 Case2E[ξ1]≤4and E[ξ1?ξt1]>4??ηt: Booking methods Acceptable classes through time t Acceptable classes thereafter DLP without re-solving1,21,2 DLP with re-solving1,21 Case3E[ξ1]>4and E[ξ1?ξt1]≤4??ηt: Booking methods Acceptable classes through time t Acceptable classes thereafter DLP without re-solving11 DLP with re-solving11,2 Case4E[ξ1]≤4and E[ξ1?ξt1]≤4??ηt: Booking methods Acceptable classes through time t Acceptable classes thereafter DLP without re-solving1,21,2 DLP with re-solving1,21,2 Suppose now the demand for the whole horizon has the following distribution: ξ1= 5with probability1 2 , 1with probability1 2 , ξ2= 2with probability1 2 , 4with probability1 2 . Moreover,suppose that we can divide the time horizon into two periods such that in the ?rst period there are two class-2arrivals only.The capacity is c=4.Notice that we have E[ξ1]=3<4andξt1=0w.p.1,so E[ξ1?ξt1]=3. From the tables above we see that the initial policy determined by the bid prices is to accept all the requests.Then,after the?rst period,two seats are occupied,i.e.,?ηt=2,and thus case2above applies.It follows that,when re-solving model[DLP]to obtain new bid prices,the policy becomes only accepting class-1customers.It is not dif?cult to verify that the expected revenue for the second period is$155.95under non-re-solving policy,and$150 under the re-solving one.Since the expected revenue for the?rst period is the same for both policies,it follows that the re-solving policy behaves worse than the non-resolving one. Similar results are obtained for the case of bid prices generated from model[SLP](under the same demand distribution as above),although the calculations are slightly more compli-cated.The solution of the SLP problem is(2,2)T,which implies that the incremental price for the leg is$60.Thus,in the?rst time period two seats are allocated to the class-2cus-tomers,so the remaining capacity for the second period is2.When re-solving SLP model again,we get the new allocation at(2,0)T which implies that the incremental price for the leg is$100.That is,the re-solving SLP method changes policy from accepting all customers into accepting only class-1customers.It follows that the expected revenue for the second period is the same as in the DLP case—$155.95for non-re-solving,$150for re-solving. This shows that re-solving can be worse under the bid-price policy,even if it is generated from the SLP model. 5Choosing when to re-solve An issue that does not seem to have been given attention in the literature is how to choose the times at which decisions are re-evaluated.The standard practice,both for multi-stage models as well as for re-solving approaches,appears to be to review the decisions at equally-spaced time points.However,there is no reason this must be done so,and one may bene?t from a better choice of those times.In this section we discuss this issue in the context of the re-solving approach laid out in Sect.3. To illustrate,consider a situation where re-solving is applied once,at some time t.That is,we have an initial allocation x0and a revised allocation x1,which is obtained from the problem solved at time https://www.doczj.com/doc/633246601.html,ing the notation de?ned earlier,letξ1andξ2be the vectors of total number of requests during intervals(0,t](stage1)and(t+1,τ](stage2)respectively. We shall writeξ1(t)andξ2(t)to emphasize the dependence of these quantities on t,and similarly for x1.Notice that,under this allocation policy,the expected revenue from time t on is given by f T E[min{x1(t),ξ2(t)}].Thus,the improvement from re-solving is given by f T E min{x1(t),ξ2(t)} ?f T E min{(x0?min{x0,ξ1(t)}),ξ2(t)} ≥0.(10) The second term above discounts the revenue resulting from keeping policy x0from time t on.The term min{x0,ξ1(t)}gives the number of sold seats up to time t,so x0minus that quantity is the number of available seats at time t under the initial policy.Thus,in principle one could determine the value of t that maximizes the quantity in(10);but doing so,of course,is not practical,as it requires re-solving the model for each value of t is order to calculate x1(t). Nevertheless,we can study a heuristic approach to determine appropriate re-solving points.We assume throughout this section that the arrival process of each itinerary-class is a(possibly non-homogeneous)Poisson process,soξ1(t)andξ2(t)are independent for any given t.Let us consider a simpli?ed model which is restricted to one leg and where the random variables in the objective function are replaced with their expectations.At time0 we solve the problem max f T min{x,E[ξ]}: n k=1 x k≤c,x≥0 (11) and the solution gives the initial allocation x0.Assume that the classes are ordered such that f1≥f2≥···≥f n.Then,we can easily solve(11).The solution is x0= E[ξ1],...,E[ξq],c? q k=1 E[ξk],0,...,0 , where q is the smallest index such that q+1 k=1 E[ξk]>c.For simplicity,assume that q k=1 E[ξk]=c,so x0q+1=0. Now consider the re-solving problem at time t.Letξ1(t)andξ2(t)be de?ned as before, soξ1(t)+ξ2(t)=ξ.Then,the re-solving problem is max f T min{x,E[ξ2(t)]} subject to n k=1 x k≤c? n k=1 min{x0 k ,?ξ1k(t)},x≥0. Note that the above problem is de?ned for a particular realization?ξ1k(t)ofξ1k(t).From the value of x0calculated above,we can rewrite the problem as max f T min{x,E[ξ?ξ1(t)]} subject to n k=1 x k≤c? q k=1 min{E[ξk],?ξ1k(t)},x≥0. (12) Again this is easy to solve,and we get x1(t)= E[ξ1?ξ11(t)],...,E[ξ ?ξ1 (t)],c? k=1 E[ξk?ξ1k(t)],0,...,0 , where is the smallest index such that +1 k=1E[ξk?ξ1k(t)]>c? q k=1 min{E[ξk],?ξ1k(t)}. Again,for simplicity let us discard the“residual”,i.e.,assume that x1 +1(t)=0.Also,sup- pose that =q w.p.1.Then,the expected objective function value of problem(12)(calcu- lated over the possible realizations of ξ1(t))is given by ν1 (t)=E f T min {x 1(t),E [ξ?ξ1(t)]} =q k =1 f k E [ξk ?ξ1k (t)].Next,we calculate the value of re-solving.If we do not re-solve,the allocation at time t is given by ˉx(t)=x 0?min {x 0,ξ1(t)},which given the value of x 0calculated above can be written as ˉx(t)= E [ξ1]??ξ 11(t) +,..., E [ξq ]??ξ1q (t) +,0,...,0 .The value of re-solving is given by the expected improvement in the objective value of problem (12)when using the optimal solution x 1(t)instead of keeping ˉx(t) .The expected objective value at ˉx(t)is given by ˉν(t)=E f T min {ˉx(t),E [ξ?ξ1(t)]} =q k =1 f k E min [E [ξk ]?ξ1k (t)]+,E [ξk ?ξ1k (t)] .In order to compare ν1(t)and ˉν (t),de?ne (t):=ν1(t)?ˉν(t).Then,we have (t)=q k =1 f k E [ξk ?ξ1k (t)]?E min [E [ξk ]?ξ1k (t)]+,E [ξk ?ξ1k (t)] .(13) To alleviate the notation,de?ne μk :=E [ξk ],and let Z k :=ξ1k (t)(so λk :=E Z k ≤μk ).Note that the second term in (13)can be rewritten as E min [μk ?Z k ]+,μk ?λk =E min {μk ?min {μk ,Z k },μk ?λk } =μk ?E max {min {μk ,Z k },λk } ,so by substituting the above into the expression for (t)we have (t)=q k =1 f k (E [max {min {μk ,Z k },λk }]?λk )=q k =1f k E [[min {μk ,Z k }?λk ]+].(14) Note that we can write the expectation inside the sum as E [min {μk ,Z k }?λk ]+ =E [min {μk ,Z k }?λk ]+I {Z k >μk } +E [min {μk ,Z k }?λk ]+I {Z k ≤μk } =E (μk ?λk )I {Z k >μk } +E [Z k ?λk ]+I {λk ≤Z k ≤μk } =(μk ?λk )P (Z k >μk )+E Z k I {λk ≤Z k ≤μk } ?λk P (λk ≤Z k ≤μk ).(15) For simplicity,let us assume that μk and λk are integers.Then,since Z k has Poisson distri-bution (with parameter λk = t 0λk (s)ds ,where λk (·)is the arrival rate function),it is easy to show that E Z k I {λk ≤Z k ≤μk } =λk P (λk ?1≤Z k ≤μk ?1) so in(14)we have (t)= q k=1 f k[(μk?λk)P(Z k>μk)+λk(P(Z k=λk?1)?P(Z k=μk))].(16) Clearly,all terms on the right-hand side of the above equation can be evaluated numerically for a given value of t,so it is easy to?nd the re-solving point t that maximizes (t). However,we are interested in developing a heuristic that allows for multiple re-solving points.To do so,we shall replace the probabilities in(16)with quantities that are easier to calculate. Consider the term P(Z k>μk).Using the Markov inequality P(Z k>μk)≤E Z k/μk,we will replace that quantity withλk/μk,so the?rst term inside the brackets in(16)becomes simplyλk?λ2k/μk.Consider now the term P(Z k=λk?1).Since Z k is a Poisson random variable with meanλk,we have P(Z k=λk?1)=e?λkλλk?1 k (λk?1)! . Using Sterling’s approximation n!≈n n e?n √ 2πn and also the approximation [n/(n?1)]n≈e,we have that P(Z k=λk?1)≈1 λk λk?1 2π , and so we see that the second term in(16)is less than or equal to λk?1 2π .It follows that the contribution of this term is small compared toλk?λ2k/μk and hence we will discard it. Next,suppose that we can actually compute different re-solving points for each class.Then, the problem of maximizing (t)in(14)becomes separable and,together with the above approximations,it reduces to?ndingλk that maximizes f k(λk?λ2k/μk).It is easy to check that the solution to this problem isλk=μk/2.That is,we want to?nd t k such that E[ξ1k(t k)]=E[ξk]/2.(17) This suggests that,for the original non-separable problem of maximizing (t)in(16),we choose t such that i f k E[ξ1k(t)]≈ 1 2 i f k E[ξk].(18) We can give the following intuitive explanation for(18).From(10)we see that,ifξ2(t)is small with high probability,the improvement in revenue resulting from re-solving is min-imal.That is,we want to pick a re-solving point t in such a way that the demand from time t on(i.e.ξ2(t))is“high enough”.This suggests taking t not too close toτ.On the other hand,if t is too close to zero,thenξ1(t)is small and x0is close to x1,so again from (10)we see that the improvement is minimal.So,it seems sensible to choose t such that E[ξ1(t)]≈E[ξ2(t)],i.e.,E[ξ1(t)]≈E[ξ]/2.We then weight the terms by the correspond-ing fares to take revenue into account. The above heuristic can be generalized to multiple re-solving points.Suppose we decide we want to re-solve R times.Then the time we pick for the r th re-solving point is t such that k f k E [ξk (t)]≈r R +1 k f k E [ξk ].(19) We omit the superscript 1from ξk (t),which represents class-k demand up to time t .Notice that ξk =ξk (τ).One way to pick t such that (19)holds is to solve the one-dimensional problem min t | k f k E [ξk (t)]?r/(R +1) k f k E [ξk ]|.The above procedure is intuitive in the single leg case,since there is a natural ordering of the classes by fare.However,as mentioned earlier the situation is more complicated when dealing with a network environment.For example,consider a situation where the fare for a certain itinerary-class pair that uses legs 1and 2is $130,whereas the fare for another itinerary-class pair that uses leg 1only is $100,and the fare for another itinerary-class pair that uses leg 2only is $70.Intuitively,if one expects to see more arrivals of the latter classes,then the second class should be preferred over the ?rst one when making decisions about leg 1.One way to quantify this is through the net contribution of each class.Suppose for example that the bid price associated with each leg is $40.Then,the net contribution of the ?rst itinerary-class is $130?$80=$50,whereas the net contribution of the second one is $100?$40=$60.That is,the second class is more pro?table even though its fare is smaller. It is clear that,in a single-leg environment,the net contribution is actually the fare level.For networks,one can apply heuristic procedures to rank the classes based on the net con-tribution.Such procedures are proposed in Bertsimas and De Boer (2005)and de Boer et al.(2002),for example.Borrowing from their ideas,we apply the follow steps to aggregate the demand vector into a one-dimensional quantity.Let f jk be the fare level for certain itinerary-class pair (j,k)and let S jk denote the set of legs which are used for that itinerary.We use the following algorithm to determine R re-solving points: Algorithm 1 1.Solve model [SLP]at time 0.Let p i denote the bid price for leg i ,obtained from the dual variables of the continuous relaxation of [SLP]. 2.Let q jk be the net contribution of itinerary-class (j,k),calculated as f jk ? i ∈S jk p i . 3.For each r ,r =1,...,R ,the r th re-solving point t r is a t that satis?es j,k q jk E [ξjk (t)]≈r R +1 j,k q jk E [ξjk ],where as before ξjk (t)is the class-k demand for itinerary j up to time t . It is worth pointing out that,when the arrival processes {ξjk (t)}are homogeneous Pois-son processes,we have E [ξjk (t)]=λjk t for all t and hence the above algorithm yields t r :=rτ/(R +1),i.e.,equally-spaced time points—which seems reasonable given the ho-mogeneity of the arrivals.As we shall see in the next section,however,in the absence of homogeneity the choice of re-solving times becomes very important. 6Numerical results In this section we describe the results from numerical experiments performed with the poli-cies discussed above.Although our data set was randomly generated,we tried to mimic real data as much as possible.To do so we imposed the following features,which according to Weatherford et al.(1993)are characteristic of actual booking processes.They are (1)un-certain number of potential customers;(2)uncertain mix of high-and low-fare customers; (3)uncertain order of arrivals;and (4)high-fare customers tend to arrive after the low-fare ones. The ?rst example is a 10-leg network described in Fig.1.We consider all ?ights to/from the hub from/to each city,as well as the ?ights between two cities connecting at the hub.Therefore,there are 30possible itineraries in the network.There are two booking classes for each ?ight,with the proportion of 1:3between high and low fare classes in terms of total requests.Following Weatherford et al.(1993),we model the booking process by a doubly stochastic non-homogeneous Poisson process (NHPP),where the arrival intensity at time t has gamma distribution.More speci?cally,for each itinerary j let λj 1(t)and λj 2(t)be the arrival intensity of respectively high-fare and low-fare customers at time t .Denote by αj >0the total expected number of requests for itinerary j over the booking horizon (i.e.,for both classes together).Let G j be a random variable with gamma distribution with shape parameter αj and scale parameter β =1(that is,the density function of G j is f j (x)=(x/β )αj ?1e ?x/β j ,x ≥0). We de?ne λjk (t),k =1,2,as λjk (t)=βjk (t)×G j ×ψk where βjk (t)=1τ t τ a jk ?1 1?t τ b jk ?1 (a jk +b jk ) (a jk ) (b jk ) .The parameters ψ1,ψ2are set with the goal of re?ecting the proportion of arrivals for high-and low-fare customers.We take this proportion to be 1:3in all itineraries,so we set ψ1=0.25and ψ2=0.75.Notice that,for each t ,λjk (t)has gamma distribution with shape parameter αj and scale parameter βjk (t)ψk .In particular,E [λjk (t)]=αj βjk (t)ψk and hence the total expected number of arrivals for itinerary j is τ0E [λj 1(t)]+E [λj 2(t)]dt =αj ψ1+αj ψ2=αj ,which is consistent with our de?nition of αj . The parameters βjk (t)are selected to re?ect the arrival patterns of different classes.High-fare customers tend to arrive close to the end of the booking horizon,whereas low-fare customers usually appear early in the booking process.To model that,we set a j 1>b j 1(high-fare customers),and a j 2 1 Fig.2Example 2 high and low fares are respectively f j 1=$500and f j 2=$100.For one-leg itineraries,we set the total expected number of requests equal to 40,that is,αj =40,with the high and low fares set as f j 1=$300and f j 2=$80.All legs in the network have capacity equal to 400,and the booking horizon has length τ=1000time units. The second example is depicted in Fig.2.Again,we consider two classes for each itinerary.Notice that there are 10one-leg,12two-leg,and 8three-leg itineraries.The ex-pected number of requests are 60,150,and 100respectively.The fare levels for different type of itineraries are set as ($300,$80),($500,$100),and ($700,$200).The parameters a jk ,b jk ,ψ1and ψ2are the same as in the ?rst example,as well as the horizon length.The leg capacities are 400for the arcs connecting the hubs with the satellite nodes,and 1,000for the arcs connecting the two hubs. For each of the problems we implemented four basic policies:DLP,SLP,DLP-based bid price and SLP-based bid price.The linear (integer)programs required to determine the allocations were solved using the software package XPress MP T M from Dash Optimization (under the Academic Partnership Program).For each of the policies,we considered the effect of solving it only once as well as twice and ?ve times over the booking horizon.The re-solving points were calculated using two methods—the standard approach of equally-spaced points as well as Algorithm 1from Sect.5. For these examples,?nding the appropriate re-solving points according to Algorithm 1was very simple,since the demand distributions used in the examples allow us to calculate E [ξjk (t)]exactly.It turns out that E [ξjk (t)]=αj ψk B jk (t/τ),where B j 1(s):=?6s 7+7s 6, B j 2(s):=?6s 7+35s 6?84s 5+105s 4?70s 3+21s 2 for all j .Notice that the expected number of arrivals of itinerary-class (j,k)over the whole horizon is E [ξjk ]=E [ξjk (τ)]=αj ψk B jk (1)=αj ψk .Figure 3shows the plot of the function H (t):= j,k q jk E [ξjk (t)]for both examples.It is clear that re-solving occurs more often as the slope of H gets larger;thus,equally-sized intervals are appropriate only when H is linear (which,as mentioned earlier,happens when the arrival process is a homogeneous Poisson),but this is not the case in these examples.The algorithm suggests re-solving more often as the end of the horizon approaches,which re?ects the fact that high-fare customers tend to book later than low-fare ones. The results in Tables 1–2show the average revenue for each policy and each example.These numbers were obtained by building a simulation model whereby we simulate the arrival process and apply the corresponding booking policies.The accrued revenue is the sum of the fares of accepted customers,and we compute the average revenue over 1000replications.Next to each number we display the half-width of a 95%con?dence interval for 一、主要设备技术参数 1.报警主机技术参数型号：LHD6000-４F 简介：有线无线兼容，通过电话通讯网络向用户传递报警信息，并可实现远程控制，对警情进行及时处理 ? 可接４路可编程有线防区，４路可编程无线防 区 ? 可接１个键盘,每个键盘可扩展2路可编程有 线防区 ? 可以硬件恢复出厂用户密码，可以软件恢复系 统出厂值 ? 可以使用２个遥控器进行现场布/撤防 ? 通过外部电话拨打报警系统，实现异地布/撤防 ? 发生警情时，自动拨打报警中心或者个人手机、固定电话进行语音报警 ? 配置灵活：可选用防火、防盗探头、玻璃破碎感应器等配件 ? 主机防拆功能、键盘防拆功能、键盘通信线路防剪功能 ? 记录最近发生的40条报警事件信息，可随时查询 ? 兼容安定宝通信协议，可联网报警，也可单机单户使用 ? 3组定时自动布/撤防时间设置 ? 交流掉电、电池欠压、电话掉线等故障报警? 内置实时时钟? 用户地址录音功 ? 主机外壳尺寸: 274mm×264mm×86mm ? 键盘尺寸: 150mm×95mm×28mm ? 额定工作电流: 交流220V ? 静态工作电流: 120mA ? 后备电池: 12V,7AH密闭铅酸蓄电池 ? 警铃输出: 10.5-13.5VDC/0.5A ? 工作温度: -10℃～+55℃ ? 遥控器有效控制距离: 无障碍物阻挡时可达100米3.红外线探测器技术参数 功能描述 产品主要性能： 1、双红外通道检测，智能波形处理，误报率 低 2、热敏电阻温度补偿 3、防破坏 4、四档位灵敏度设置 5、指示灯开关设置 6、有线标准接口输出 7、可与各种报警控制器配套使用 产品技术参数 1、产品尺寸：长×宽×高 100×62×45mm 2、包装重量：135g 各种消防报警模块的参数、特点及使用方法 1特点 GST-LD-8321中继模块采用DC24V供电，总线信号输入与输出间电气隔离，完成了探测器总线的信号隔离传输，可增强整个系统的抗干扰能力，并且具有扩展探测器总线通讯距离的功能。GST-LD-8321中继模块主要用于总线处在有比较强的电磁干扰的区域及总线长度超过1000m需要延长总线通讯距离的场合。 2主要技术指标 （1）总线输入距离≤1000m （2）总线输出距离≤1000m （3）电源电压：DC18V～DC27V （4）静态工作电流≤20mA （5）带载能力及兼容性：可配接1～242点总线设备，兼容所有总线设备 （6）隔离电压：总线输入与总线输出间隔离电压≥1500V （7）使用环境： 温度：-10℃～+50℃ 相对湿度≤95％，不结露 （8）外形尺寸： 85mm×128mm×56mm 采用隔离方式进行总线信号传输，安装在总线上，用于总线长度超过1000米时扩展总线距离，或现场存在强电磁干扰时进行总线隔离。 1特点 GST-LD-8300型输入模块用于接收消防联动设备输入的常开或常闭开关量信号，并将联动信息传回火灾报警控制器（联动型）。主要用于配接现场各种主动型设备如水流指示器、压力开关、位置开关、信号阀及能够送回开关信号的外部联动设备等。这些设备动作后，输出的动作信号可由模块通过信号二总线送入火灾报警控制器，产生报警，并可通过火灾报警控制器来联动其它相关设备动作。输入端具有检线功能，可现场设为常闭检线、常开检线输入，应与无源触点连接。本模块可采用电子编码器完成编码设置。当模块本身出现故障时，控制器将产生报警并可将故障模块的相关信息显示出来。 2主要技术指标 （1）工作电压：总线24V （2）工作电流≤1mA （3）线制：与控制器的信号二总线连接 （4）出厂设置：常开检线方式 （5）使用环境： 温度：-10℃～+55℃ 相对湿度≤95％，不结露 （6）外壳防护等级：IP30 （7）外形尺寸： 86mm×86mm×43mm（带底壳） 电子编码，可接收设备常开或常闭开关量信号。 1、 FE100接警主机 描述 FE100是Honeywell的另一款报警接收机，它结构紧凑，功能强大，且价廉物美。FE100采用DSP数字信号处理技术，使得其通信环境适应能力极强。FE100完全兼容ADEMCO和原C＆K 报警主机的通讯格式，还有来电显示等功能。 FE100的内置软件设置方便，用户可据自己的实际需要对FE100作相应的设置变更，从而达到与新产品兼容。 实践证明，FE100是一款非常适合国内组建报警中心的报警接收机。 功能： ?可接入8条电话线，同时处理8个用户的报警 ?来电显示 ?内置电话线错误检测 ?抓擢功能(需与软件及MODEM配合) ?兼容市场上几乎所有流行的通讯格式，即可接收大多数品牌的报警主机的报警信号 ?多达5000条的事件记录 ?内置完善的抗雷击功能及噪音过滤功能 ?设计精巧，功耗极小 特性： ?电压：220 V AC，50 Hz ?后备电池：12VDC，6AH ?通讯协议(部分)： ·ADEMCO Contact lD格式 ·l El fast格式 ·4+2(1400Hz)格式 ·4+2(2300Hz)格式 ·C＆K格式 ·C＆K(bell)格式 输出： ·TerminaI(终端) ·打印机 ·RS-232 尺寸：400X200X353mm 2、 Monitor XP 7.0 标准版 Monitor XP7.0监察者系列报警中心管理软件是北京迈特安技术发展有限公司最新开发的新一代联网报警中心管理软件。经过了近十年的积累和完善，已经被广泛应用于传统的公用电话网（PSTN）联网报警中心和新型的网络化联网报警中心。目前已有包括专业保安公司、公安、部队、金融、通信等行业在内的数千家不同规模的报警中心在应用Monitor XP 软件，在行业相关产品的市场占有率为80%以上，成为各保安协会推荐的以及各报警中心建设、改造的首选专业软件。 功能特点： ?可容纳上千用户入网 海湾公司消防设备技术性能阐述 一、海湾报警设备应用介绍及主要设备功能简要 系统功能 海湾公司提供的火灾自动报警及消防联动控制系统，可以实现以下控制功能： a、智能电源盘包括直流电源、浮充备用电源及电源监控装置三个部分，电源监控部分采用CPU数码显示电压及电流，用来指示正在使用电源的输出电压值及输出电流值，以及各类故障及状态显示。 b、能控制常用电梯，使其自动降至首层，接收反馈信号并显示状态。 c、火灾时能输出警报装置投入工作的控制信号。 d、火灾时能将火灾疏散层的扬声器和公共广播扩音机强制转入火灾应急广播状态。并将反馈信号显示。 e、火灾时能输出控制的疏散、诱导指示设备投入工作的控制信号。 f、火灾时能对失火区域疏散通道上的门禁系统控制器进行解锁门禁，并将反馈信号显示。 g、能在管网气体灭火系统的报警、喷洒各个阶段发出相应的声、光警报信号，声信号能手动清除；在延时阶段能输出关闭相关的防火门、窗，停止空调通风系统，关闭相关部位防火阀的控制信号，接收反馈信号并显示状态。 h、能停止有关部位的空调通风、关闭电动防火阀的控制信号，接收反馈信号并显示状态 i、通过设置编码输入输出模块能控制用作防火间隔的防火卷帘门的控制信号，接收反馈信号并显示状态。 j、通过设置编码双输入双输出模块能控制疏散通道防火卷帘门的半降、全降的控制信号，接收反馈信号并显示状态。 k、能启动有关部位的防烟、排烟风机和排烟阀等的控制信号，并接收反馈信号并显示状态，排烟风机能自动、手动或手动直接控制启动。 l、可通过自动、手动或手动直接控制消防水泵的启动和停动，接收反馈信号并显示状态。消火栓手动报警按钮带编码可显示其所在位置。 火灾自动报警系统技术规格书 1.系统概述 1.1 系统构成 在让胡路西新建信号楼，还建列检所信息机房及还建驼峰信号楼，分别设置区域火灾自动报警系统，在24小时有人值班室内设置区域火灾报警控制器，在设备用房设置气体灭火控制盘、感烟探测器、感温探测器，并设置监视模块、控制模块、声光报警器、放气指示灯等。 1.2 电源 在让胡路西新建信号楼，还建列检所信息机房及建驼峰信号楼区域火灾报警控制器附近设置1套24V消防电源。 1.3 设备配置需求表 序号物资设备名称规格型号单位数量 1火灾报警控制器套3 2气体灭火控制盘套3 3消防电源套3 4智能型感烟探测器个7 5智能型感温探测器个13 6声光报警器个10 7放气指示灯个5 8紧急启停按钮个5 9监视模块个4 10控制模块个3 2．系统功能 2.1消防控制设备对气体灭火系统有下列控制、显示功能 显示系统的手动、自动工作状态； 在报警、喷射各阶段，控制室有相应的声、光警报信号，并能手动切除声响信号； 在延时阶段，自动关闭防火门、窗，停止通风空调系统，关闭有关部位防火阀； 显示气体灭火系统防护区的报警、喷放及防火门（帘）、通风空调等设备的状态。 2.2火灾报警后，消防控制设备对防烟、排烟设施有下列控制、显示功能： 停止有关部位的空调送风，关闭电动防火阀，并接收其反馈信号； 启动有关部位的防烟和排烟风机、排烟阀等，并接收其反馈信号； 控制挡烟垂壁等防烟设施。 （1）本招标文件用户需求及要求，主要由各相关的消防工程设计图纸和相应的用户需求要求组成。 （2）投标人应对招标文件中的技术要求及项目逐项答复,并应进行必要的说明。与招标技术文件有差异的地方应列出差异表，并做详细说明。 （3）如投标人没有以书面形式对本需求书提出异议，则意味着投标人所提供的设备完全符合本需求书的要求，如有异议，投标人应在投标书中以“对需求书的意见和同需求书的差异”为题的专门章节中加以详细描述。 （4）在正式合同签订前，招标方保留对本招标文件进行解释的权利。合同签订后，招标方保留对本招标文件解释的权利，遇有修改，双方协商解决。 （5）投标人须对所提供的产品的质量和售后服务做出承诺。提出系统在质保期内的服务承诺及系统在质保期后的维护计划和维护方案。 （6）本技术标书及要求是最低限度的技术要求，并未对一切技术细节做出规定，也未充分引述有关标准和规范的条文，投标方应保证提供消防报警控制器、气体灭火控制器及联动模块是符合本用户需求书的要求和有关工业标准的进口产品，获得UL及FM认证，且为同一厂家系列标准产品。 （7）本技术标书及要求所使用的标准和规范如与投标方所执行的标准发生矛盾时，应按较高标准执行。 （8）投标方所提供的系统和货物，如若发生侵犯专利权的行为时，其侵权责任与招标人无关，应由投标方承担相应的责任，并不得损害招标方的利益。 （9）投标方应仔细阅读招标文件的全部条文，对于招标文件中存在的任何含糊、遗漏、相互矛盾之处或是对于用户需求以及其它内容不清楚、认为存在歧视、限制的情况，投标方应在规定时间之前向招标人寻求澄清。 （10）由于相关专业的设计还未稳定，投标方应充分考虑本项目的未确定因素进行投标方案设计，任何未确定因素引起的变化将不影响总价。 （1）投标方所提供的设备必须是信誉可靠、技术先进、且有成熟的运用实例。 （2）FAS系统中所使用的各种火灾探测器和火灾报警控制器等火灾报警产品需获得中国消防产品质量认证委员会颁发的产品质量认证证书，并经大庆市公安消防局或哈尔滨铁路局公安消防处备案登记。产品必须在明显位置粘贴中国消防产品质量认证委员会印制的安全认证标志。 （3）系统的主要组件（包括报警控制器、气体灭火控制器，点式感烟、感温探测器等）由同一厂商供应，并采用国际国内知名品牌设备。 （4）系统的设备，包括安装中所使用的设备、材料、布线方法、安装工艺、调试开通及验收等，均应符合国家的有关规范及标准。 （5）本条款仅列出主要设备的要求，其它附件及材料须符合中国有关标准并经业主认可方可使用。 （6）投标方提供的产品如非本厂生产，应提供中国消防产品质量认证委员会颁发的产品质量认证证书。 （7）投标方应在不增加价格条件下，提供供货时的主流电子产品。 （8）FAS系统不能因单点设备故障（包括但不限于开路、短路及接地），影响整个系统的正常运转。 第三部分项目采购内容 一、项目概况 1.1、项目名称：奉贤区卫生服务机构管理中心社区卫生安防报警系统设备采购项目1.2、采购编号：12-00067 1.3、采购预算：350000元（最高限价） 二、采购设备及设备要求 1、实时图像监控系统 1.1系统技术指标： 复合视频信号幅度：1.0Vp-p±3dB 同步信号幅度：0.3Vp-p±3dB 彩色水平清晰度：480线 信噪比：不低于37dB 图像等级：不低于4级 灰度等级：不低于8级 1.2主要设备技术指标： （1）摄像机 成像器件：1/3英寸CCD； 有效像素：44万，PAL752(水平) X 582(垂直)；NTSC768(水平) X 494(垂直) 水平清晰度：480TVL； 最低照度：彩色：0.1Lx(F1.2 ,30IRE ,AGC ON)；支持红外光辅助照明； 信噪比：不低于48dB； 电子快门范围：1/50到1/100000秒； 工作温度：-10℃～＋60℃； 电源： 12VDC。 （2）硬盘录像机 视频压缩标准：H.264； 视频处理芯片：Davinci处理器； 操作系统：嵌入式操作系统； 视频分辨率：CIF/QCIF； 视频输入:16路，BNC接口，1.0Vp-p,75Ω 视频输出：2路，BNC接口，1.0Vp-p,75Ω，1路VGA输出 通讯接口：1个10M/100M自适应以太网口，1个RS232口，1个RS485口； 硬盘应采用串行接口：支持8个1000G的SATA硬盘； 电源电压：AC 220V 工作温度：-10℃～＋55℃ 1.3系统工作原理 （1）前端彩色摄像机的视频信号经过对应的视频输入口进入硬盘录象机进行录影及保存。 （2）在硬盘录像机中，配置1块1000G的硬盘，可对5路图象进行存储，保存30天以上的图像资料，便于资料的取证和查询。硬盘录像机还配置了网络接口，可以经网络在远端通过软件有权限地观看实时图像，并可下载机内录像资料，进行调看。 火灾探测器智能光电感烟 智能感温 普通光电感烟 普通感温 独立光电感烟 模块类输入模块 输出模块 中继模块 隔离模块 火灾报警设备火灾报警控制器（联动型） 火灾报警控制器（联动型） 火灾显示盘 火灾报警控制器 图形显示装置 消防联动控制设 备 立柜式 琴台式 按钮产品火灾报警按钮 消火栓按钮 声光产品声光报警器 气体灭火系统气体灭火控制器 防火卷帘系统进入 可燃气体报警 系统可燃气体报警控制器 可燃气体报警探测器 智能喷灌控制器 电气火灾监控设 备 进入 J TY-GD-OT502智能光电感烟火灾探测器 简介 OT502智能光电感烟探测器采用红外散射原理设计，内置先进的MCU微处理器，具有强大的现场数据采集和分析判断能力。完善的火灾判定智能模型和漂移自动补偿技术，有效确保了探测器报警的准确性和工作的稳定性。 获得了公安部消防产品合格评定中心颁发的3C质量认证证书(新标准) 通过了国家消防电子产品质量监督检验中心检验 功能 采用高性能微处理器，内含Flash存储器，功能强大，性能可靠。 内置优置红外光电管，内含日光滤波器，有效滤除环境光线干扰。 特殊的光学迷宫结构，响应快，一致性好。 先进的自动编址功能，节约工程施工时间。 环境自动补偿,独特的自诊断技术，对环境变化（温度、湿度、灰尘污染）的漂移量具有自动修正补偿功能，极大地降低了系统的误报率。 内含智能软件，与控制器双向分布智能，最大限度地保证了报警的准确性。 二总线电流量脉宽数字化信号传输技术，通讯可靠，抗干扰性能强。 总线无极性，避免了由于接线不当而引起的系统损坏。 具有过流保护功能，可监测其供电电压，并可在控制器上显示出来，方便工程调试。 先进的SMT贴片工艺、可靠的屏蔽措施，对电磁环境恶劣场所有很强的抗干扰能力。 特殊三防处理,防霉防潮防腐蚀，抗潮湿性能强。 平时绿灯闪亮，报火警时红灯常亮。 超薄流线外形，美观大方。 技术指标 工作电压：DC20～30V; 监视电流：＜350uA; 报警电流：<28mA; 风速：＜10 m/s; 保护面积：60～80m2，详见《火灾自动报警系统设计规范》; 环境条件温度：-10℃～＋50℃ 湿度：≤95％（40±2℃）; 编码方式：自动编码或手动编码，范围1～192; 线制：两总线，无极性; 适用底座：ODZ5004A、ODZ5006A. 火灾自动报警设备技术参数-江苏赛福特电子一、火灾探测器 1(SF5111探测器底座 2(JTY-GD-SF5131点型光电感烟火灾探测器 3(SF5131-EX防爆点型光电感烟火灾探测器 4(JTW-ZD-SF5151点型感温火灾探测器(A2) 5(SF5151-EX防爆点型感温火灾探测器 6(JTW-ZD-SF5151/C点型感温火灾探测器(A2) 7(JQB-HX2131B 编址型可燃气体探测器 8(防爆气体探测器 9(JTY-H-VDC1382A线型光束感烟探测器 二、手动火灾报警按钮 1(J-SAP-M-SF5143手动火灾报警按钮 2(J-SAP-M-SF5143/P手动火灾报警按钮(带电话插孔) 3(SF5143-EX防爆手动火灾报警按钮 三、模块类 1(SF5147编码型信号输入模块 2(SF5146 多功能输入编码接口 3(SF5149编码型总线输出模块 4(SF5145 总线广播模块 5(SF5141 联动双切换模块 6(SF5139总线隔离模块 7(SF5148/B编址型声光报警器 四、控制器 1(SF5100系列控制器 2(JB-TB-SF5100火灾报警控制器 3(JB-TG-SF5100 火灾报警控制器 3(1 SF5100/AB320 显示盘 3(2 SF5100/LA 回路监控单元 3(3 SF5100/CD8多线控制盘 3(4 SF5100/CK50总线控制盘 4(控制器组网 4(1控制器对等组网 4(2控制器主从方式组网 4(3无线组网 4(4远程通信 5(JB-QB-SF5000区域火灾报警控制器五、气体灭火系统 1(SF5100/CE4 气体灭火控制盘 2(JB-QB-SF5100/CE2气体灭火控制器3(SF5171 气体灭火控制模块 4(SF5172紧急启停按钮 5(SF5148非编址火灾声光报警器 6(SF5173气体释放显示灯 六、配套产品 1(SF5181楼层火灾显示盘 2. SF5182液晶楼层火灾显示盘 （一）火灾报警控制器 ● JB-QT-GST9000型火灾报警控制器（联动型） 设计采用JB-QT-GST9000型火 灾报警控制器（联动型）作为控制中 心报警控制器，其主要特点如下： 采用10英寸彩色液晶显示屏， 各种报警状态信息均可以直观的以 汉字方式显示在屏幕上，便于用户操 作使用；可以观察智能型探测器动态 工作曲线，以便于了解现场的实际环 境条件；控制器具有强大的面板控制及操作功能，各种功能设置全面、简单、方便。 控制器采用模块化设计，具有高度智能化的特点，与智能探测器一起可组成分布智能式火灾报警系统，系统工作可靠性高，极大地降低了误报。 系统内部采用并行总线数据传输，主控板与各回路板之间信息传输采用查询的工作方式，不会因为多回路而影响整机的巡检速度，信息传输快速、准确。 全总线控制系统布线灵活，既可以采用报警点与联动点共回路布线方式，布线简单；也可以采用采用报警点与联动点分回路布线方式，系统工作可靠性高。 控制器具有极强的现场编程能力，各回路设备间的交叉联动、各种汉字信息注释、总线制控制设备与多线制控制设备之间的相互联动等均可以现场编程设定。 控制器可完成自动及手动控制外接消防被控设备，其中手动控制 方式具备直接手动操作键控制输出及编码组合键手动控制输出二种方式，系统内的任一地址编码点既可由各种编码探测器占用，也可由各类编码模块占用，设计灵活方便。联动控制设备采用专用24V 直流电源供电，使联动设备故障不会影响到主机的正常工作。 主要技术指标如下： 液晶屏规格：10英寸彩色液晶屏，640×480图形点阵 ◎控制器容量： a.最多可带58个 242地址编码点回路，最大容量为14000个地址编码点 b.可外接128台火灾显示盘；联网时最多可接32台其它类型控制器 c.多线制控制点及直接手动操作总线制控制点可按要求配置 ◎线制： a.控制器与探测器间采用无极性信号二总线连接，与各类控制模块间除无极性二总线外，还需外加二根DC24V 电源总线。 b.与其它类型的控制器采用有极性二总线连接，对于火灾报警显示盘，需外加两根DC24V 电源供电总线。 c.与彩色CRT 系统采用四芯扁平电话线，通过RS-232标准接口连接，最大连接线长度不宜超过15m 。 ◎使用环境：温度： 0℃～＋40℃ 相对湿度≤95％，不结露 ◎电源：主电为交流220V +10%-15%；控制器备电为DC24V/24Ah 密封铅电池； 联动备电为DC24V/24Ah 密封铅 电池。 ◎控制器监控功耗＜150W ；控制器最大功耗＜250W 。 ◎外形尺寸：1045mm ×933mm ×1350mm 海湾公司火灾自动报警系统主要设备技术特点及参数 2009年8月版 前言 本技术文件详细介绍了海湾公司生产销售的民用及工业场所类火灾自动报警及消防联动控制主要设备的技术特点及参数。涵盖了火灾报警控制器、火灾探测器、报警按钮、现场模块及接口设备、指示部件、火灾报警显示盘、GST-GM9000图形显示系统、气体灭火控制系统、消防电话系统、消防广播系统等系列产品的相关内容。 本文内容全面详实，图文并茂，可作为火灾自动报警及消防联动控制产品的选型及应用设计的参考资料使用。刷绿色产品，暂未取得检验报告。 目录 （一）火灾报警控制器 (1) ● JB-QB-GST100型火灾报警控制器 (1) ● JB-QB-GST200型火灾报警控制器 (3) ● JB-QB-GST200型火灾报警控制器（联动型） (5) ● JB-QB-GST500型火灾报警控制器（联动型） (7) ● JB-QG-GST5000型火灾报警控制器（联动型） (9) ● JB-QT-GST5000型火灾报警控制器（联动型） (11) ● JB-QG-GST9000型火灾报警控制器（联动型） (13) ● JB-QT-GST9000型火灾报警控制器（联动型） (16) （二）火灾探测器 (19) 1、编码型火灾探测器 (19) ● JTY-GD-G3型点型光电感烟火灾探测器 (19) ● JTY-GM-GST9611型点型光电感烟火灾探测器 (21) ● JTW-ZCD-G3N型点型差定温火灾探测器 (23) ● JTY-ZOM-GST9612型点型差定温火灾探测器 (24) ● JTF-GOM-GST601型点型复合式感烟感温火灾探测器 (26) ● JTF-GOM-GST9613型点型复合式感烟感温火灾探测器 (27) ● JTY-HM-GST102型线型光束感烟火灾探测器 (29) ● JTG-ZW-G1型点型紫外火焰探测器 (31) ● JTG-ZM-GST9624型点型紫外火焰探测器 (33) 2、非编码型火灾探测器 (34) ● JTY-GF-GST104型点型光电感烟火灾探测器 (34) ● JTY-GF-GST9711型点型光电感烟火灾探测器 (35) ● JTWB-ZCD-G1（A）型点型差定温火灾探测器 (36) ● JTW-ZOF-GST9712型点型差定温火灾探测器 (37) ● JTFB-GOF-GST601型点型复合式感烟感温火灾探测器 (38) ● JTFB-GOF-GST9713型点型复合式感烟感温火灾探测器 (39) ● JTY-HF-GST102型线型光束感烟火灾探测器 (41) ● JTG-ZW-G1B型点型紫外火焰探测器 (43) 主要设备技术参数要求 一．UPS参数要求： 1、本次UPS系统供货范围:双变换在线式工频UPS，容量为50KVA、100KVA，采用工频双变换技术，内置输出隔离变压器，自带静态旁路和市电维修旁路开关；内置独立的主路输入开关、旁路开关和输出开关，确保后级设备和人身安全。 2、UPS主机须具备共用电池组的功能。 3、远程监控通信接口应有RS232，RS485，可选干结点接口，SNMP 卡，并提供相应通信协议并负责与用户的监控软件对接联通、实现正常的远程监控功能。 4、主机须采用先进的DSP数字处理电路，使UPS系统超稳定运行。智能侦测系统的微处理器不间断地对所有的电源状态、断路器状态、熔断器状态和所有的电路工作状态进行在线侦测。出现故障时，侦测系统会即时报警通知管理员，同步启动UPS全面保护功能。 5、采用全数字化控制技术，实现8台UPS并联冗余功能。取消传统的插件式电路处理工艺，全部采用高精度SMD技术，采用4层电路板设计和高精度SMD元件完全清除由芯片自身产生的各种高频信号对其他芯片的干扰，从而让各个芯片模块能够不受干扰的正常工作，便于提高集成电路的安全运行可靠性和运行精度。 6、电池智能化管理；可根据用户电池配置自动调整电池的充电电流参数，并根据供电环境对电池进行均充浮充转换、温度补偿充电和放电管理。（可选配电池巡检仪）对每节电池都必须进行在线检测，能预测电池组的剩余运行时间，可选短信报警器即发生故障时可无线向指定手机和远程监控系统报警。 7、功率逆变器必须采用第六代IGBT模块，具有更低的饱和压降，逆变器的工作效率更高，温升低，可靠性更高。 8、*UPS主机可根据用户用电要求对UPS进行工作状态设置，用户可选UPS工作模式、ECO工作模式、EPS工作模式（提供第三证明文件加盖公章的复印件） 9、UPS主机具有远程电话网络语音监控系统功能 目录 （一）火灾报警控制器 (1) ● JB-QT-GST9000型火灾报警控制器（联动型） (1) ● JB-QG-GST9000型火灾报警控制器（联动型） (3) ● JB-QT-GST5000型火灾报警控制器（联动型） (5) ● JB-QG-GST5000型火灾报警控制器（联动型） (7) ● JB-QB-GST500型火灾报警控制器（联动型） (9) ● JB-QG-GST200型火灾报警控制器（联动型） (11) ● JB-QB-GST200型火灾报警控制器（联动型） (13) （二）火灾探测器 (15) ● JTY-GD-G3型点型光电感烟火灾探测器 (15) ● JTW-ZCD-G3N型点型差定温火灾探测器 (17) ● JTY-GF-GST104型点型光电感烟火灾探测器 (18) ● JTWB-ZCD-G1（A）型点型差定温火灾探测器 (19) ● JTY-HM-GST102型线型光束感烟火灾探测器 (19) ● JTG-ZW-G1型点型紫外火焰探测器 (21) ● JTF-GOM-GST601型点型复合式感烟感温火灾探测器 (22) （三）报警按钮 (23) ● J-SAP-8401型手动火灾报警按钮 (23) ● J-SAP-8402型手动火灾报警按钮 (24) ● LD-8403型消火栓按钮 (25) ● GST-LD-8404型智能编码消火栓报警按钮 (26) （四）防爆系统产品 (27) ● JTY-GF-GST104（Ex）本安型光电感烟探测器 (27) ● JTWB-ZCD-G1（A）（Ex）本安型电子差定温感温探测器 (28) ● JTFB-GOF-GST601（Ex）本安型烟温复合探测器 (29) ● J-SAB-G1（Ex）型手动火灾报警按钮 (30) ● LDB-8403（Ex）本安型消火栓报警按钮 (31) 1、 FE100接警主机描述 FE100是Honeywell的另一款报警接收机，它结构紧凑，功能强大，且价廉物美。FE100采用DSP数字信号处理技术，使得其通信环境适应能力极强。FE100完全兼容ADEMCO和原C＆K 报警主机的通讯格式，还有来电显示等功能。 FE100的内置软件设置方便，用户可据自己的实际需要对FE100作相应的设置变更，从而达到与新产品兼容。 实践证明，FE100是一款非常适合国内组建报警中心的报警接收机。 功能： ?可接入8条电话线，同时处理8个用户的报警 ?来电显示 ?内置电话线错误检测 ?抓擢功能(需与软件及MODEM配合) ?兼容市场上几乎所有流行的通讯格式，即可接收大多数品牌的报警主机的报警信号 ?多达5000条的事件记录 ?内置完善的抗雷击功能及噪音过滤功能 ?设计精巧，功耗极小 特性： ?电压：220 V AC，50 Hz ?后备电池：12VDC，6AH ?通讯协议(部分)： ·ADEMCO Contact lD格式 ·l El fast格式 ·4+2(1400Hz)格式 ·4+2(2300Hz)格式 ·C＆K格式 ·C＆K(bell)格式 输出： ·TerminaI(终端) ·打印机 ·RS-232 尺寸：400X200X353mm 2、Monitor XP 7.0 规范版 Monitor XP7.0监察者系列报警中心经管软件是北京迈特安技术发展有限公司最新开发的新一代联网报警中心经管软件。经过了近十年的积累和完善，已经被广泛应用于传统的公用电话网（PSTN）联网报警中心和新型的网络化联网报警中心。目前已有包括专业保安公司、公安、部队、金融、通信等行业在内的数千家不同规模的报警中心在应用Monitor XP 软件，在行业相关产品的市场占有率为80%以上，成为各保安协会推荐的以及各报警中心建设、改造的首选专业软件。 功能特点： ?可容纳上千用户入网 ?全面兼容CFSK III、Contact ID、4+2 Express等协议 ?支持多种联网方式 ?支持多种接警设备 ?数据库稳定可靠 ?二次开发方便 ?操作员分级经管 ?操作简单直观 ?声光电报警提示 ?详尽的报告查询与统计 ?用户设置报告触发功能 ?全局和多级电子地图功能 ?多级联网报警转发功能 ?预置处警预案功能 ?单据经管功能 ?报表统计功能 ?资料导出功能 ?计划任务功能 ?事件查询功能 海康威视50米红外摄像机技术参数 主要特性 ?采用高性能SONY CCD ?分辨率高,600TVL,图像清晰、细腻 ?低照度,0.1Lux @ (F1.2,AGC ON),0 LUX with IR ?支持自动彩转黑功能,实现昼夜监控 ?符合IP66级防水设计,可靠性高 海康威视30米红外摄像机技术参数 主要特性 采用高性能SONY CCD ?分辨率高,600TVL,图像清晰、细腻 ?低照度,0.1Lux @ (F1.2,AGC ON),0 LUX with IR ?支持自动彩转黑功能,实现昼夜监控 ?符合IP66级防水设计,可靠性高 海康威视高清D1录像机参数 海康威视红外高速云台参数 主要特性红外功能： ? 最低照度0Lux ? 采用高效红外阵列，低功耗，照射距离达80m ? 红外灯与倍率距离匹配算法，根据倍率及距离调节红外灯亮度，使图像达到理想的状态 ? 内置热处理装置，降低球机内腔温度，防止球机内罩起雾 ? 恒流电路设计，红外灯寿命达3万小时 系统功能： ? 采用1/4"索尼高性能CCD, 图像清晰 ? 精密电机驱动, 反应灵敏, 运转平稳, 精度偏差少于0.1度, 在任何速度下图像无抖动 ? 支持RS-485控制下对HIKVISION、Pelco-P/D协议的自动识别 ? 支持三维智能定位功能, 配合DVR和客户端软件可实现点击跟踪和放大 ? 支持多语言菜单及操作提示功能, 用户界面友好 ? 支持数据断电不丢失 ? 支持断电状态记忆功能, 上电后自动回到断电前的云台和镜头状态 ? 支持光纤模块接入 ? 支持内置温度感应器, 可显示机内温度 ? 支持防雷、防浪涌、防突波功能 ? 室外球达到IP66防护等级 ? 支持RS-485线路故障诊断功能, 把故障信息, 如地址错误、波特率错误等以文字形式显示在视频画面上 ? 支持曼码协议及线路故障诊断功能, 把故障信息, 如地址错误、波特率错误等以特殊字符形式显示在视频画面上 ? 支持定时启动预置点/花样扫描/巡航扫描/水平扫描/垂直扫描/随机扫描/帧扫描/全景扫描等功能 ? 支持密码保护功能, 防止被人恶意修改球机菜单参数 ? 支持球机标题功能, 可在视频画面叠加中、英文字符 ? 支持区域扫描和显示, 球机在设定的区域设定的时间内没收到控制命令就执行区域扫描, 并显示区域名称 机芯功能： ? 支持自动光圈、自动聚焦、自动白平衡、背光补偿和低照度(彩色/黑白)自动/手动转换功能, 宽动态功能可选 ? 可设置多达24块隐私屏蔽区域, 位置、大小可调整, 画面内可同时有8块区域被屏蔽 云台功能： ? 水平方向360°连续旋转, 垂直方向-10°-90°, 支持自动翻转, 无监视盲区 ? 水平预置点速度最高可达240°/s, 垂直预置点速度最高可达200°/s ? 水平键控速度为0.1°~160°/s, 垂直键控速度为0.1°~120°/s ? 支持256个预置位, 并具有预置点视频冻结功能 ? 支持8条巡航扫描, 每条可添加32个预置点 ? 支持4条花样扫描, 总记录时间大于10分钟 ? 支持比例变倍功能, 旋转速度可以根据镜头变倍倍数自动调整 **项目 小区智能化系统主要设备选型及技术参数 二、小区智能化系统主要设备技术参数 （一）、硬盘录像机技术参数 BST-8000-16D嵌入式硬盘录像机(D1格式) 功能特点： 1、采用最新的USB2.0接口，进一步稳定USB鼠标功能，快速实现USB备份、USB升级等操作； 2、采用高性能DSP实现标准的H.264压缩算法，图像更精细、码流更小、网传更流畅； 3、采用标准网络协议和标准压缩算法，在各种平台上轻松实现互连互通； 4、采用SATA硬盘接口、支持SATA刻录备份； 5、采用双码流技术，更适合窄带传输，录像支持实时； 6、D1/HD1/CIF/QCIF优化组合； 7、适用于画质和网络传输要求较高、存储时间较长的金融、能源、城市治安等监控场合 技术参数： 型号：BST-8000-12D 主处理器：工业级嵌入式微控制器 操作系统：嵌入式LINUX操作系统 系统资源：同时多路录像，同时录像回放，同时网络操作 操作界面：16位真彩色图形化菜单操作界面，支持鼠标操作，带有菜单注释 画面显示：1/4/8/9/13/16画面显示 视频标准：PAL(625线，50场/秒） 监视图像质量：PAL制，D1(704×576) 回放图像质量：PAL制、QCIF(176×144)、CIF(352×288)、BCIF(528×288)、 HD1(352×576)、D1(704×576) 图像压缩：H.264限定码流、H.264可变码流 图像控制：6档可调 录像速度(CIF) ：PAL制：每4路200fps自由组合 图像移动侦测：每画面可设置192(16×12)个检测区域；可设置多级灵敏度 录像方式及优先级：手动>报警>动态检测>定时 本地回放：支持2路回放 录像查询方式：时间点检索、日历检索、事件检索、通道检索 每路占用硬盘空间：56～900M字节/小时 报警系统设备参数 目 录 目录 1.控制主机 (5) 1.14110DL控制主机 (5) 1.2VISTA-10P控制主机 (6) 1.3VISTA-20P控制主机 (7) 1.4VISTA-120/250总线制大型控制主机 (9) 1.5Vista-128BPT/ Vista-250BPT (11) 1.6COMPACT-4 4防区控制主机【GPRS主机】 (12) 1.7236 PLUSII控制主机 (13) 1.8238C PLUSＩＩ控制主机 (14) 1.9238C Super 控制主机 (15) 1.102316PLUSＩＩ控制主机 (18) 1.112316super 控制主机 (19) 1.12L YNX家居无线控制主机 (21) 2.报警模块及附件 (24) 2.1报警键盘 (24) 2.1.16148CH固定字符键盘 (24) 2.1.26160可编程英文液晶键盘 (24) 2.1.3236PLUS LED控制键盘 (25) 2.1.4238CPLUS LED控制键盘 (25) 2.1.52316PLUS LED控制键盘 (26) 2.1.62300Alpha Plus II 控制键盘 (26) 2.2VISTA系列附件 (27) 2.2.14229八防区扩展模块 (27) 2.2.24219八防区扩展模块 (27) 2.2.34293SN单防区扩展模块 (28) 2.2.44193SN双防区扩展模块 (28) 2.2.54193SNP双防区扩展模块 (29) 2.2.64208SN八防区扩展模块 (29) 2.2.74101SN总线继电器模块 (30) 2.2.84204 四路继电器联动模块 (30) 2.2.94232AP 32路继电器模块 (30) 2.2.104286语音模块 (32) 2.2.114100SM串行接口模块 (32) 2.2.12IP-2000网络接口展模块 (33) 2.2.13IPM-VISTA网络接口模块 (33) 2.2.14IPM-VISTA super II网络接口模块 (34) 2.2.154297总线延伸模块 (35) 2.2.16VSI总线隔离器 (36) 2.2.17VISTA-KEY 门禁控制器模块 (36) 2.2.185881 ENH 防区无线接收机 (37) 标准文档 报警系统设备参数 目 录 目录 1.控制主机 (5) 1.14110DL控制主机 (5) 1.2VISTA-10P控制主机 (6) 1.3VISTA-20P控制主机 (7) 1.4VISTA-120/250总线制大型控制主机 (9) 1.5Vista-128BPT/ Vista-250BPT (11) 1.6COMPACT-4 4防区控制主机【GPRS主机】 (12) 1.7236 PLUSII控制主机 (13) 1.8238C PLUSＩＩ控制主机 (14) 1.9238C Super 控制主机 (15) 1.102316 PLUSＩＩ控制主机 (18) 1.112316 super 控制主机 (19) 1.12LYNX家居无线控制主机 (21) 2.报警模块及附件 (24) 2.1报警键盘 (24) 2.1.16148CH固定字符键盘 (24) 2.1.26160可编程英文液晶键盘 (24) 2.1.3236PLUS LED控制键盘 (25) 2.1.4238CPLUS LED控制键盘 (25) 2.1.52316PLUS LED控制键盘 (26) 2.1.62300Alpha Plus II 控制键盘 (26) 2.2VISTA系列附件 (27) 2.2.14229八防区扩展模块 (27) 2.2.24219八防区扩展模块 (27) 2.2.34293SN单防区扩展模块 (28) 2.2.44193SN双防区扩展模块 (28) 2.2.54193SNP双防区扩展模块 (29) 2.2.64208SN八防区扩展模块 (29) 2.2.74101SN总线继电器模块 (30) 2.2.84204 四路继电器联动模块 (30) 2.2.94232AP 32路继电器模块 (30) 2.2.104286语音模块 (31) 2.2.114100SM串行接口模块 (32) 2.2.12IP-2000网络接口展模块 (32) 2.2.13IPM-VISTA网络接口模块 (33) 2.2.14IPM-VISTA super II网络接口模块 (34) 2.2.154297总线延伸模块 (35) 2.2.16VSI总线隔离器 (35) 2.2.17VISTA-KEY 门禁控制器模块 (36) 2.2.185881 ENH 防区无线接收机 (37) 湿式报警阀结构与工作原理 一、概述： ZSFZ湿式报警阀是只允许水单方向流入喷水灭火系统，并在规定的压力和流量下驱动配套部件报警的一种单向阀。它与水流指示器、压力开关、洒水喷头等组成的湿式自动喷水灭火系统是一种应用极为广泛的固定式灭火系统。该系统管网内常年充满一定压力的清水，长期处于伺应工作状态，当保护区域内某处发生火灾时，区域内环境温度升高，洒水喷头的热敏感元件(玻璃球)中的有机溶液发生热膨胀而产生很大的内压力，直到玻璃球外壳发生破碎，从而开启喷头喷水，并且自动启动整个系统，发出声光报警信号，以达到火灾报警及控制火灾、扑灭火灾之目的。 ZSFZ型湿式报警阀装置适用于环境温度为4℃～70℃的场所。通常安装于宾馆、商场、医院、电影院、办公楼、高层建筑、仓库和地下车库等有火灾危险的场所，对于自动喷水灭火系统的消防工程设计、施工、监理请按GB50084-2001(2005)《自动喷水灭火系统设计规范》内各条款执行。 产品标准：GB5135·2-2003《自动喷水灭火系统第2部份：湿式报警阀、延迟器、水力警铃》 二、湿式报警阀结构和工作原理： 图1：报警阀结构示意图 2.1 报警阀的结构： 从图1可知，ZSFZ湿式报警阀装置由湿式报警阀、延迟器、水力警铃、压力开关、排水阀、过滤器等组成。 2.1.1、湿式报警阀 本湿式报警阀为盖板型报警阀，主要由阀体、座圈和阀瓣三部分组成，整个阀体被阀瓣分成上、下两腔，上腔(系统侧)与系统管网相通，下腔(供水侧)与水源相通，在阀体中配有座圈，在座圈上，有多个通往延迟器进水管的沟槽小孔。 当系统处于伺应状态时，座圈上的沟槽小孔被阀瓣盖住封闭，通往水力警铃的报警水道被堵死；当上、下压力差达到一定数值，阀瓣才开启(差压启动)，水就从供水侧流向系统侧，警铃报警，灭水系统喷水；当上、下压力频繁开启，在阀瓣上设有小补水阀，当系统侧管网有微小渗漏或水源压力有波动时，可以通过补水阀给管网补水，平衡上、下腔压力，稳定了阀瓣，从而避免了误报警。 2.1.2、延迟器 延迟器是一个有进水口和出水口的圆筒形储水容器，下端有进水口，与报警阀的报警口连接相通，上端有出水口，连接水力警铃。由于系统的供水源压力存在波动现象，能使阀瓣出现瞬间开启，水流经过座圈上沟槽及小孔首先进入延迟器，由于水源压力波动的时间很短，阀瓣很快就能自动复位(关闭)，所以进入延迟器的水量很小，可以由延迟器来收集水，并经过底部的泄水口排泄，延迟器的这一缓冲时间作用，避免了水流压力波动而引起水力警铃的误报警。水从延迟器的进水口流入到出水口流出所需要时间为延迟时间，本装置为5~90S，水流停止后，遗留在延迟器中的水由泄水口排出，排完所需时间(排水时间)小于5min。 2.1.3、水力警铃 水力警铃是一种水力驱动的机械装置。由壳体、叶轮、铃锤和铃盖等组成。当阀瓣被打开，水流通过座圈上的沟槽和小孔进入延迟器，充满后，继续流向水力警铃的进水口，在一定的水流压力下，推动叶轮带动铃锤转臂旋转，使铃锤连续击打铝铃而发出报警铃声。 2.2 工作原理 湿式报警阀装置长期处于伺应状态，系统侧充满工作压力的水，自动喷水灭火系统控制区内发生火警时，系统管网上的闭式洒水喷头中的热敏感元件受热爆破自动喷水，湿式报警阀系统侧压力下降，在压差的作用下，阀瓣自动开启，供水侧的水流入系统侧对管网进补水，整个管网处于自动喷水灭火状态。同时，少部份水通过座圈上的小孔流向延迟器和水力警铃，在一定压力和流量的情况下，水力警铃发出报警声响，压力开关将压力信号转换成电信号，启动消防水泵和辅助灭火设备进行补水灭火，装有水流指示器的管网也随之动作，输出电信号，使系统控制终端及时发现火灾发生的区域,达到自动喷水灭火和报警的目的.报警主机技术参数型号：LHD6000-4F全解
各种消防报警模块的参数、特点及使用方法
防盗报警设备技术参数.doc
海湾消防设备技术性能阐述
火灾自动报警系统技术规格书
安防报警系统技术参数与投标文件格式
火灾自动报警设备技术参数-深圳奥瑞那
火灾自动报警设备技术参数
火灾报警系统主要设备的技术特点及参数
火灾自动报警系统主要设备的技术特点及参数(新国标8月版)
主要设备技术参数要求
火灾报警系统主要设备技术特点及参数
防盗报警设备技术参数论述
监控设备详细技术参数
小区智能系统主要设备品牌选型及技术参数
霍尼韦尔报警系统设备参数版
霍尼韦尔报警系统设备全参数(2015版)
湿式报警阀结构工作原理和技术参数
相关主题
文本预览