Demeulemeester, E.L., Herroelen, W.S.: A Branch-and-bound Procedure for the Multiple Resource-Constrained Project Scheduling Problem. Management Science 38(12), pp. 1803-1818. (1992)  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... not incorporated in our branch and bound procedure all the results obtained by other researchers for the Resource Constrained Project Scheduling Problem (RCPSP) In particular, we have not used, until now, any intelligent backtracking rule such as the cut set rule of Demeulemeester and Herroelen [15]. This may seem a little strange given the excellent results reported in [16] in particular on the KSD instances, even with a limited use of the cut set rule. However, it appears that many industrial scheduling problems include several additional features (including, for example, elastic ....

E. Demeulemeester and W. Herroelen, A Branch-and-Bound Procedure for the Multiple Resource-Constrained Project Scheduling Problem, Management Science 38 (1992) 1803-1818.

....problem (RCPSP) is known as an NP hard problem. Therefore the main focus is on the development of branch and bound algorithms where different ideas have been presented to built the tree guiding the enumeration of the schedules. The schemes proposed enumerate, e.g. delaying alternatives (cf. [4]) extension alternatives (cf. 16] feasible posets (cf. 11] feasible sequences (cf. 12] 13] and feasible subsets (cf. 10] in order to find a schedule with a minimum makespan. The currently most advanced procedure has been developed by Demeulemeester and Herroelen (cf. 6] It builds on ....

....(cf. 12] 13] and feasible subsets (cf. 10] in order to find a schedule with a minimum makespan. The currently most advanced procedure has been developed by Demeulemeester and Herroelen (cf. 6] It builds on ideas from Christofides et al. cf. 2] and enhances their earlier work (cf. [4]) by a bound introduced by Mingozzi et al. cf. 10] and the full exploitation of nowadays available 32 bit architecture of personal computers. The procedure has solved the entire set of benchmark problems generated by ProGen (cf. 9] for the first time. The projects consist of 32 activities ....

[Article contains additional citation context not shown here]

Demeulemeester, E. L. and W.S. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....modern concepts include more realistic aspects. In general, the availabilities of the resources involved are limited. Consequently, numerous publications have dealt with exact methods for solving the so called singlemode resource constrained project scheduling problem (SRCPSP) cf. e.g. 4] 5] [6], 17] 27] where each of the activities comprising the project has to be performed in one prescribed way (mode) using certain amounts of the resources provided. The objective considered is, as performed by CPM and MPM, the minimization of the makespan. Recent advances have incorporated a ....

....Example Network For ease of notation we focus on the single mode case and consider the example given in Figure 1 (cf. 9] p. 179) Obviously activities 2, 3 and 4 cannot be started before activity 1 is finished. If activity 1 is scheduled activity 2, 3 and 4 become eligible. In contrast to e.g. [6], 25] an activity is 7 called eligible if all its predecessors are scheduled but not necessarily finished. Using graph theoretical terminology the activities 2, 3 and 4 will be denoted as descendents (sons) of activity 1 (the father) These relationships are depicted by the precendence tree ....

[Article contains additional citation context not shown here]

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....and RS 0.2, 0.5, 0.7, 1.0 . For each instance cluster defined by a combination of these parameters, J30 contains ten instances, for a total of 480 instances. Of these, those 120 instances where RS = 1.0 are trivially solvable by the MPM schedule. Of the remaining sample, the exact algorithm of Demeulemeester, Herroelen (1992) found and verified the optima for 308 instances within a time limit of 3600 seconds per instance, taking on average 615.1 seconds per instance on a 386SX 15 PC; ten more instances were optimally solved but not verified within the time limit. These 308 instances have also been considered in ....

DEMEULEMEESTER, E. AND W.S. HERROELEN (1992), "A branch-and-bound procedure for the multiple resourceconstrained project scheduling problem", Management Science 38, pp. 1803-1818.

....general, depth rst search is used in order to keep the memory requirements low. Di erent methods using di erent branching schemes and pruning methods have been developed: Patterson et al. 1989] and Sprecher [1994] use the concept of a precedence tree, the methods of Christo des et al. 1987] and Demeulemeester Herroelen [1992] are based on delay alternatives, while Stinson et al. 1978] use extension alternatives. Mingozzi et al. 1998] use a slightly di erent approach. 3.4 Heuristics for the RCPSP To solve an RCPSP of medium to large size in reasonable time one has to apply heuristic methods. The following types of ....

Demeulemeester, E., Herroelen, W. [1992] A branch-and-bound procedure for the multiple resource-constrained project scheduling problem, Management Science 38. 1803-1818.

.... t = ST g i 1 1; T ; e) ST prec g i 1 (PS j ) ST g i 1 then, by the use of the scheduling strategy given in Theorem 1, Part I, PS i Phi [g i 1 ; m g i 1 ] is dominated by PS j Phi [g i 1 ; m g i 1 ] Proof: Obvious 2 However, obviously, in contrast to the single mode case (cf. e.g. [4]) in the multi mode case condition (d) of Remark 2 cannot be deduced from (d1) CT g CT g ; if g 2 CS(PS i ) with CT g ST g i 1 (d2) CT g ST g i 1 ; otherwise: That is, in order to prevent from excessive use of storage and expensively proving the assumptions of Remark 2 we strengthened them to ....

....multi mode resource constrained project scheduling problem. The size of the projects that can be solved to optimality has been nearly doubled. However, although the rules presented can mainly be implemented in other multi mode suitable generalizations (cf. e.g. 17] of single mode algorithms (cf. [4], 20] the approach presented here remains the most general one currently available. In contrast to the previously mentioned it covers time varying resource availabilities and can simply be generalized to cover time varying resources requests as well. By the use of the standard project generator ....

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....because of mode coupling via resource constraints. 5 Computational Results 5. 1 Single Mode Case Currently the most advanced exact procedure for solving makespan minimization problems seems to be the implicit enumeration procedure of the B B type with backtracking from Demeulemeester (cf. 17] [18]) It is coded in C and solves the fourty three 27 job problems out of the 110 Patterson instances in an average computational time of 1.06 seconds to optimality on an IBM PS 2 Model 55sx (80386sx processor, 15 Mhz clockpulse) We used the original implementation of the algorithm provided by ....

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, to appear.

....scheduling problem (RCPSP) is known as an NP hard problem (cf. 6] Therefore the main focus is on the development of branch and bound algorithms where different ideas have been presented to build the tree guiding the enumeration of the schedules. The schemes enumerate delaying alternatives (cf. [3], 4] feasible completion times (cf. 19] feasible extensions (cf. 18] feasible posets (cf. 13] feasible sequences (cf. 12] 14] 15] and feasible subsets (cf. 11] in order to find an optimal, i.e. makespan minimal, solution. The currently most studied procedure has been developed ....

....[19] feasible extensions (cf. 18] feasible posets (cf. 13] feasible sequences (cf. 12] 14] 15] and feasible subsets (cf. 11] in order to find an optimal, i.e. makespan minimal, solution. The currently most studied procedure has been developed by Demeulemeester and Herroelen (cf. [3], 4] In the major part of business applications one can employ different resources and or quantities to define alternative ways, i.e. modes, to execute the activities that comprise a project. The activity duration is a discrete function of the employed quantities. That is, the concept allows ....

[Article contains additional citation context not shown here]

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resourceconstrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

.... 1996) indeed, assuming a deadline to be given for the project completion even its feasibility variant is strongly NP complete (Garey, Johnson 1975) Exact algorithms include implicit enumeration methods (Talbot, Patterson 1978; Christofides et al. 1987; Alvarez Vald s, Tamarit 1989; Tavares 1990; Demeulemeester, Herroelen 1992, 1995; Mingozzi et al. 1994; Sprecher 1996) zero one programming (Bowman 1959; Pritsker et al. 1969; Patterson, Huber 1974; Patterson, Roth 1976) and dynamic programming (Carruthers, Battersby 1966) Yet, b)owing to the complexity of the problem, most algorithms devised for it are heuristic ....

....measured between minimum and maximum demand. For each combination of these parameters, J30 contains ten instances, for a total of 480 instances. Of these, the 120 instances where RS = 1.0 are trivially solvable by the MPM schedule. Of the remaining nontrivial sample, the exact algorithm of Demeulemeester, Herroelen (1992) found and verified the optimal solutions for 308 instances within a time limit of 3600 seconds per instance, taking on average 615.1 seconds per instance on a 386SX 15 PC; ten more instances were optimally solved but not verified within the time limit. These 308 instances have also been ....

DEMEULEMEESTER, E. AND W.S. HERROELEN (1992), "A branch-and-bound procedure for the multiple resourceconstrained project scheduling problem", Management Science 38, pp. 1803-1818.

....resource constrained project scheduling problem (SMRCPSP) where each of the activities of the project has to be performed in one prescribed way (mode) using specified amounts of the resources provided. The most common objective of the SMRCPSP is the minimization of the makespan (cf. e.g. [3, 4, 5, 24, 32]) Recent developments have incorporated more reality by allowing the activities to be executed in one out of several modes. The modes reflect alternative combinations of resources and belonging quantities employed to fulfill the tasks related to the activities. The activity duration is a discrete ....

....[1. 10] MM, as well as additionally generated problem sets have been used in numerous publications. In the following we give a brief summary. 5. 1 Single Mode Instances Kolisch et al. 17, 18] solved the instance set J30 with the exact solution procedure of Demeulemeester and Herroelen (cf. [5]) for studying the influence of the variation of project characteristics, like the number of activities J , the number of renewable resources jRj, the resource factor RFR , the resource strength RSR , and the network complexity NC, on the computation time of the exact branch andbound procedure. As ....

[Article contains additional citation context not shown here]

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, No. 12, pp. 1803-1818.

....(1996) that for R 2 this algorithm may not find the optimal solution and, moreover, that for instances with N 1, the algorithm may even be unable to determine an existing feasible solution. Finally, Sprecher et al. 1997) extended the concept of delaying alternatives as introduced by Demeulemeester Herroelen (1992) for the single mode RCPSP to mode and delayingalternatives. Their implementation solves the 10 activity instances in an average of 0.53 CPUseconds on an IBM compatible 80386dx personal computer with 40 MHz clockpulse. Despite this encouraging results it has to be recalled that exact algorithms ....

Demeulemeester, E. and W.S. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem, Management Science, Vol. 38, pp. 1803-1818.

....project as an instance of the well known resource constrained project scheduling problem (RCPSP) which additionally covers resource availability and request varying with time. For the classical RCPSP, the currently most powerful exact algorithms of Brucker et al. 2] Demeulemeester and Herroelen [4, 5], Mingozzi et al. 19] and Sprecher [22] are able to solve instances with 30 activities and 4 renewable resources, whereas problems with 60 activities and more are still intractable if the resources are scarce. Resource requirements that vary with time, as to be considered for the research ....

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....the former of the multi mode case of project scheduling, the latter of all types of scarce resources, containing renewable and nonrenewable resources as special cases. 2 Handbook on Recent Advances in Project Scheduling Traditional resource constrained project scheduling approaches (cp. e.g. Demeulemeester and Herroelen 1992, Patterson 1984) have been restricted to the case where each activity may be performed in only one predefined way. In this context, the resource constrained project scheduling problem (RCPSP) is one of the most important problems (cp. e.g. Baar et al. 1997, Brucker et al. 1998, Demeulemeester ....

Demeulemeester, E., and W. Herroelen. 1992. A branch--and--bound procedure for the multiple resource--constrained project scheduling problem. Mgmt. Sci. 38, 1803--1818.

....(de Werra 1985, Abramson 1991) course scheduling (Drexl, Juretzka, and Salewski 1993) and other assignment type scheduling problems. Traditional resource constrained project scheduling approaches (Davis and Patterson 1975, Talbot and Patterson 1978, Stinson, Davis, and Khumawala 1978, Demeulemeester and Herroelen 1992) have been restricted to the case where each job may be performed in only one predefined way. More recently efforts have been made to formulate and solve the more general preemptive project scheduling problem where job durations are functions of consumed resources (Blazewicz, Cellary, Slowinski, ....

Demeulemeester, E. and W. Herroelen, "A Branch-and-Bound Procedure for the Multiple ResourceConstrained Project Scheduling Problem", Management Science, 38 (1992), 1803-1818.

.... problem as well (cf. 9] Consequently, the RCPSP has caused numerous publications dealing with the developement of suboptimal (cf. e.g. 2] 11] and optimal (cf. 1] 3] 13] 15] 16] 21] solution procedures, where Demeulemeester and Herroelen s exact solution procedure (cf. 4] [5]) outperforms all the other approaches. By considering the activity durations as a discrete function of the resources and or amounts of the resources allocated, we obtain a more realistic model. The time resource and resource resource tradeoff can be implemented and the multi mode ....

....activity j, 1 j J , respectively. The benefit of the time windows is twofold: First, they can be used in the mathematical programming formulation to reduce the number of variables substantially. Second, they can be utilized in several enumeration procedures to speed up the convergence (cf. e.g. [5], 19] and Section 5) For the formal description of the MRCPSP we define a binary variable for each combination of an activity j, 1 j J , a mode m, 1 m M j , and a period t, t = 0; T (cf. 22] x jmt = 1; if job j is performed in mode m and completed in period t 0; otherwise. ....

[Article contains additional citation context not shown here]

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....Consequentially, if one insists on algorithms guaranteed to find an optimal solution, the current state of the art still has only exponential answers to offer. These include implicit enumeration methods (Talbot, Patterson 1978; Christofides et al. 1987; Alvarez Vald s, Tamarit 1989; Tavares 1990; Demeulemeester, Herroelen 1992, 1997; Mingozzi et al. 1994; Sprecher 1996) zero one programming (Bowman 1959; Pritsker et al. 1969; Patterson, Huber 1974; Patterson, Roth 1976) and dynamic programming (Carruthers, Battersby 1966) Yet, considering the complexity of the RCPSP, it comes as no surprise that the majority of ....

....7: KSD Instance Set J30 Varied Design Parameters For each combination of these parameters, J30 contains ten instances, for a total of 480 instances. Of these, the 120 instances where RS = 1.0 are trivially solvable by the MPM schedule. Of the remaining nontrivial sample, the exact algorithm of Demeulemeester, Herroelen (1992) found and verified the optimal solutions for 308 instances within a time limit of 3600 seconds per instance, taking on average 615.1 seconds per instance on a 386SX 15 PC; ten more instances were optimally solved but not verified within the time limit. These 308 instances have also been ....

DEMEULEMEESTER, E. AND W.S. HERROELEN (1992), "A branch-and-bound procedure for the multiple resourceconstrained project scheduling problem", Management Science 38, pp. 1803-1818.

....such that the makespan of the project is minimized. This problem arises within project management software as well as within systems for production planning and scheduling. The currently most powerful exact procedures have been presented by Brucker et al. 2] Demeulemeester and Herroelen [5, 6], Mingozzi et al. 22] and Sprecher [26] However, they are unable to find optimal schedules for highly resourceconstrained projects with 60 activities or more. Hence, in practice heuristic algorithms to generate near optimal schedules for larger projects are of special interest. Recently, an ....

....parameters are assumed to be nonnegative and integer valued. The objective is to determine a schedule with minimal makespan such that both the precedence and resource constraints are fulfilled. Mathematical programming formulations of the RCPSP have been given by e.g. Demeulemeester and Herroelen [5, 6], Mingozzi et al. 22] and Sprecher [26] 3 Permutation based Genetic Algorithm 3.1 Basic Scheme Introduced by Holland [11] genetic algorithms (GAs) serve as a heuristic meta strategy to solve hard optimization problems. For an introduction into GAs, we refer to Goldberg [7] In this section ....

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

....are assumed to be nonnegative and integer valued. The objective is to determine a schedule with minimal makespan such that both the precedence and resource constraints are fulfilled. Mathematical programming formulations of the RCPSP have been given by, e.g. Demeulemeester and Herroelen [6, 7], Mingozzi et al. 22] and Sprecher [30] 4 Self Adapting Genetic Algorithm for Project Scheduling 4.1 Basic Scheme This section introduces a self adapting GA for the RCPSP. It is based on a genotype which consists of a problem representation of so called activity lists as well as information ....

E. L. Demeulemeester and W. S. Herroelen. A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, 38:1803--1818, 1992.

.... are described by Pritsker et al. 1969) Balas (1970) Davis and Heidorn (1971) Schrage (1971) Gorenstein (1972) Fisher (1973) Patterson and Huber (1974) Patterson and Roth (1976) Stinson et al. 1978) Talbot and Patterson (1978) Christofides et al. 1987) Bell and Park (1990) and Demeulemeester and Herroelen (1992), respectively. Stinson et al. 1978) proposed a new lower bound based on the longest path of the precedence graph that takes into account some of the relaxed resource constraints. Fisher (1973) was the first to derive a lower bound based on a Lagrangean relaxation of resource constraints. A ....

....graph obtained from the precedence graph by adding disjunctive arcs to partially represent resource constraints. The branch and bound procedure based on this later approach is shown to perform better than Stinson s procedure for a number of problem instances with 25 activities and 3 resources. Demeulemeester and Herroelen (1992) describe a new effective branch and bound algorithm that outperforms all other exact methods, at least on the set of 110 test problems assembled by Patterson (1984) Kolisch et al. 1992) introduce a number of parameters to identify easy and hard RCPSP instances and generated a new set of test ....

[Article contains additional citation context not shown here]

Demeulemeester E., Herroelen W., "A Branch and Bound Procedure for the Multiple ResourceConstrained Project Scheduling Problem", Management Science, 38, 12, (1992), 1803-1818.

No context found.

Demeulemeester, E.L., Herroelen, W.S.: A Branch-and-bound Procedure for the Multiple Resource-Constrained Project Scheduling Problem. Management Science 38(12), pp. 1803-1818. (1992)

No context found.

E.L. Demeulemeester and W.S. Herroelen. A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, 38(12):1803--1818, 1992.

No context found.

Demeulemeester E.L. and W.S. Herroelen (1992): A Branch and Bound Procedure for the Multiple Resource Constrained Project Scheduling Problems. Management Science 38, 18031818.

No context found.

Demeulemeester E., Herroelen W. (1992). A branch-and-bound procedure for the multiple resource-constrained project scheduling problem, Management Science, 38, 1803-1818.

No context found.

Demeulemeester, E. and W. Herroelen (1992): A branch-and-bound procedure for the multiple resource-constrained project scheduling problem. Management Science, Vol. 38, pp. 1803-1818.

No context found.

Erik Demeulemeester and Willy Herroelen [1992]. A Branch and Bound Procedure for the Multiple Resource-Constrained Project Scheduling Problem. Management Science, 38(12):1803-1818, 1992.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC