E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job-shop scheduling. Management Science, 44(2):262--275, 1998. 28  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... for example, see the recent surveys by Blazewicz et al. 6] or Jain and Meeran [19] Within the class of local search algorithms, the strongest performers are typically derivatives of tabu search [6] 19] 34] the sole exception being the guided local search algorithm of Balas and Vazacopoulos [2]. The power of tabu search for the JSP is perhaps best illustrated by considering the computational effort required to locate optimal solutions to a notoriously difficult benchmark problem, Fisher and Thompson s infamous 10 10 instance [12] Nowicki and Smutnicki s algorithm [22] requires only ....

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for jobshop scheduling. Management Science, 44(2):262--275, 1998.

....experimentation are displayed. However, the MRE value displayed at the end of each table summarizes performance over all problems in the group. The SHKT heuristic ran for 720 minutes, with results displayed at various points in time. Table 3 also reports results obtained by Balas and Vazacopoulus [4] using their heuristic SB GLS1 and the best from among all their algorithms (BV best) Note the SHKT heuristic finds the previously unknown optimal solution for problem TA67, and a new upper bound for TA62. Table 1 Problem N x M Opt (LB UB) TSSB SHKT5min SHKT30min SHKT60min SHKT120min ....

Balas, E., A. Vazacopoulos, (1994). "Guided local search with shifting bottleneck for job shop scheduling," Tech Rep., Management Science Research Report MSRR-609, GSIA Carnegie Mellon University, Pittsburg.

....not satisfy production needs, at any point on the line, a bottleneck is created and quality suffers as workers rush to finish their assignments. Labor planning is made even more challenging by a highly customized product mix that changes daily and the resulting shifting bottleneck phenomenon [2]. Sound labor planning begins with the development of labor estimates or standards. There are three general approaches for developing standards [3] 1) estimation looking at a job and judging the time required to complete it, 2) using historical records, such as time clock data, to estimate ....

Balas, E. and Vazacopoulos, A., 1998, "Guided Local Search with Shifting Bottleneck for Job Shop Scheduling," Management Science, 44(2), 262-274.

.... lower and UB the upper bounds taken from [6, 8, 20] One of the most successful heuristics applied to job shop scheduling is the shifting bottleneck procedure developed by Adams et al. 2] We compare our results with a recent variant of this approach (BV SBP) published by Balas and Vazacopoulos [4]. Secondly, a comparison is performed with results obtained by Mattfeld [13] who uses genetic algorithms (M GA) For the results in column CS SAW the Table indicates the run time of our simulated annealing procedure in seconds on a Sun Ultra 1 170 SPARC machine. On the 20 Theta 20 problem YN1 ....

E. Balas and A. Vazacopoulos. Guided Local Search with Shifting Bottleneck for Job Shop Scheduling. Management Science, 44:262--275, 1998.

....Best results were obtained using the latter in the local search phase. Not surprisingly, ILS performed better than random restart given the same amount of time, for any choice of the embedded local search heuristic. In more recent work on the job shop scheduling problem, Balas and Vazacopoulos [4] presented a variable depth search heuristic which they called guided local search (GLS) GLS is based on the concept of neighborhood trees, proposed by the authors, where each node corresponds to a solution and the child nodes are obtained by performing an interchange on some critical arc. In ....

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job shop scheduling. Management Science, 44(2):262--275, 1998.

....that the problem structure is exploited as much as possible. For details see, e.g. the original papers by Lin and Kernighan [90, 102] Chapter 19 of [126] or [87] See also recent applications of the variable depth search such as [21, 127, 164, 165, 163] Note that a similar idea is also used in [13] and is called the guided local search; however, this is di erent from the one explained in Section 3.2. 25 It is also important to consider how to search the neighborhood e ciently. For this purpose, it is often e ective to eliminate from consideration those solutions having no possibility of ....

E. Balas and A. Vazacopoulos, \Guided local search with shifting bottleneck for job shop scheduling," Management Science, vol.44, pp.262-275, 1998.

....produce substantially better schedules than were previously computed. Currently, all of the best known algorithms use iterated local search. The current champions are a taboo search algorithm by Nowicki Smutnicki [16] and a technique called reiterated guided local search by Balas Vazacopoulos [4]. Vaessens, Aarts Lenstra [21] provide an excellent survey of recent developments in this area. There has been less progress in finding good lower bounds for the job shop scheduling problem. The job bound is the maximum total processing required for the operations on a single job. Similarly, ....

....we use Iterated Carlier Pinson as our initial lower bound and an upper bound provided by Applegate Cook s implementation of the shifting bottleneck procedure. For Double Shave, we use C P Shave as our initial lower bound and the best known upper bound, primarily due to Balas Vazacopoulos [4]. C P Shave provides a significant improvement over Iterated Carlier Pinson. The computation time for C P Shave ranged from 3 seconds for the fastest problem to 160 seconds for the slowest; the more difficult problems are generally in the one to two minute range. Double Shave is an excellent ....

E. Balas and A. Vazacopoulos. Guided Local Search with Shifting Bottleneck for Job Shop Scheduling. Management Science Research Report #MSRR-609, Graduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, PA, 1994.

....in acceptable time, a long list of heuristics were developed in the last 25 years. In 1988 ( Balas et al. 1988] published a very promising concept with the shifting bottleneck procedure. The best currently known algorithms are the tabu search algorithms (c.f. Nowicki and Smutnicki, 1996] and ([Balas and Vazacopoulos, 1995]) They solve the notorious 10x10 problem by Muth and Thompson optimally in less then 3 seconds on a modern personal computer. This paper gives parallel versions of the shifting bottleneck procedure and the tabu search algorithm of Nowicki and Smutnicki to enlarge their practical applicability. 2 ....

Balas, E. and Vazacopoulos, A. (1995). Guided local search with shifting bottleneck for job shop scheduling. Management Science Research Report, 609.

....after the other operations of its block, provided that the resulting orientation is feasible; otherwise, v is moved to the left or to the right as long as the orientation remains feasible. While the above neighborhood functions are based on adjacent interchanges or swaps, Balas and Vazacopoulos [9] propose a neighborhood function N 4 that uses reinsertions or jumps. More precisely, N 4 considers any two operations v and w on the same machine such that v occurs prior to w on a longest path. A neighbor is obtained by inserting v immediately after w or w immediately before v. Sufficient ....

....reschedule these machines one by one. Their extended procedure SB4 takes the best solution of SB3 and a variant 10 of SB3 that reverses the order of the two reoptimizations procedures: first reschedule some non critical machines, then apply regular iterative improvement. Balas and Vazacopoulos [9] propose a rather different procedure SB GLS. It reoptimizes partial schedules by applying their variable depth search algorithm GLS (see Section 6.3) for a limited number of iterations using the jump neighborhood function N 4 . The shifting bottleneck procedure and its variants have been ....

[Article contains additional citation context not shown here]

E. Balas and A. Vazacopoulos, 1994. Guided Local Search with Shifting Bottleneck for Job Shop Scheduling, Management Science Research Report #MSRR-609, Graduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, Pennsylvania.

.... those that are in front of the jobs. Some kind of opportunism is inevitable for ecient scheduling, even in static environments [5] otherwise it would be hard to cope with the inherent complexity of the problems. Successful scheduling methods, like the Shifting Bottleneck and its descendants [3] rst look and optimize for a bottleneck resource, and then tailor repeatedly the remaining part of the schedule around the xed schedule(s) As argued in [21] and [29] dynamic scheduling calls for an opportunistic approach, because the tightly coupled nature of the problem makes it hard to ....

....scheduling problem is really dynamic. Without further developments it can hardly work on problems where the number of open jobs is considerably larger then the number of machines. This is just on the contrary to classical approaches that regard problems with many more jobs than machines as easy [3]. Recently, 33] presented a exible decomposition scheme that can cover the spectrum between dynamic and static scheduling methods. Our work is closely related to the developments in the eld of holonic manufacturing systems [16] Holonic manufacturing is directly based on the idea of ....

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job shop scheduling. Management Science, 44(2):262-275, 1998.

....a good initial solution must be found. A large range of different heuristic methods for finding an initial solution have been proposed, and one of the most successful is the Shifting Bottleneck procedure by Adams et al. 1] later refined by Applegate and Cook [2] and by Balas and Vazacopoulus [3]. Since the aim of our research has been experiments with parallel implementations of Branch and Bound, we have, however, chosen to leave out a heuristic from our code. Instead, we have either used the best known solution found by other researchers as an initial solution or just assumed that we ....

E. Balas, A. Vazacopoulos, Guided Local Search with Shifting Bottleneck for Job Shop Scheduling, Management Science Report MSRR-609, Carnegie Mellon University, 1994

....in a second step we applied the procedure described at the end of the previous section to selected benchmarks, taking into account the running times of the rst step as well as the distance to the best known upper bounds. Problem n m lb ub lb #nodes CPU Sec ta03 15 15 1206 [Va95] 1218 [BaVa95] 1207 3064 142 ta04 1170 [Va95] 1175 [We95] 1171 2927 95 ta05 1210 [Va95] 1228 [We95] 1211 116434 5688 ta06 1210 [Va95] 1239 [BaVa96] 1211 218982 9218 ta07 1223 [Va95] 1228 [Ta93] 1224 132293 4357 ta08 1187 [Va95] 1217 [BaVa95] 1188 5557 198 ta09 1247 [Va95] 1274 [BaVa95] 1248 16264 610 dmu01 20 ....

.... Problem n m lb ub lb #nodes CPU Sec ta03 15 15 1206 [Va95] 1218 [BaVa95] 1207 3064 142 ta04 1170 [Va95] 1175 [We95] 1171 2927 95 ta05 1210 [Va95] 1228 [We95] 1211 116434 5688 ta06 1210 [Va95] 1239 [BaVa96] 1211 218982 9218 ta07 1223 [Va95] 1228 [Ta93] 1224 132293 4357 ta08 1187 [Va95] 1217 [BaVa95] 1188 5557 198 ta09 1247 [Va95] 1274 [BaVa95] 1248 16264 610 dmu01 20 15 2363 [DMU97] 2579 [Ja98] 2450 3279 345 dmu02 2452 [DMU97] 2716 [Ja98] 2550 107 8 dmu03 2540 [DMU97] 2765 [Ja98] 2650 466 48 dmu04 2486 [DMU97] 2672 [Ja98] 2550 22811 2043 dmu05 2654 [DMU97] 2785 [Ja98] 2700 401 33 swv03 20 ....

[Article contains additional citation context not shown here]

E. Balas, A. Vazacopoulos (1995), Guided Local Search with Shifting Bottleneck for Job Shop Scheduling, Management Science Research Report #MSRR-609(R) [revised version], Graduate School of Industrial Administration, Carnegie-Mellon University, Pittsburgh, Pennsylvania.

....effort on techniques that combine myopic problem specific methods, which apply a neighbourhood scheme, and a meta strategy which guides the search out of local optima. Such hybrid techniques are known as iterated local search algorithms or meta heuristics and the approaches of Thomsen (1997) Balas and Vazacopoulos (1998) and Jain and Meeran (1998) are currently providing the best results. A detailed review of local search neighbourhoods and meta heuristics is given in Blazewicz et al. 1996) and Jain and Meeran (1999) As algorithms for the DJS problem have improved, the neighbourhoods that they apply have placed ....

Balas, E. and Vazacopoulos, A. (1998) Guided Local Search with Shifting Bottleneck for Job-Shop Scheduling, Management Science, Feb, 44(2), 262-275.

.... Computational Study of the Core and Shell System The several core and shell combinations which have attained a good level of performance with respect to the makespans achieved and the solutions surpassed, namely cg1f7, cg3f7 as well as cg1h1, cg1g3 and cg3g1 are compared with the techniques of Balas and Vazacopoulos (1998), Thomsen (1997) Nowicki and Smutnicki (1996) which are currently the best methods available) Demirkol et al. 1997) and Mattfeld (1996) Results are given for each individual P J benchmark problem attempted by these various authors and the CPU time is specified as a computer independent ....

....for each of the techniques compared are detailed in table 4. Apart from the five CSF combinations specified above table 5 through to table 11 also detail the best results achieved from all the various core and shell combinations tested throughout this work (CSF Best) Two individual techniques of Balas and Vazacopoulos (1998) are analysed SB GLS2 and SB RGLS5. SB GLSk indicates that Guided Local Search (GLS) has been embedded into the Shifting Bottleneck Procedure (SBP) with the GLS procedure applied for k cycles, while SB RGLSk also denotes that SBP and GLS have been combined. Here R indicates that the machines are ....

[Article contains additional citation context not shown here]

Balas, E. and Vazacopoulos, A. (1998) Guided Local Search with Shifting Bottleneck for Job-Shop Scheduling, Management Science, Feb, 44(2), 262-275.

....on the same machine. As a result Dauz re P r s and Lasserre (1993) propose a heuristic strategy while Dauz re P r s (1995) and Balas et al. Deterministic Job Shop Scheduling: Past, Present and Future 03 10 98 page(10) 1995) utilise an exact scheme to deal with DPCs. The most recent work (Balas and Vazacopoulos 1998) amalgamates guided variable search with SBP producing one of the best P J approaches available. Eventhough Balas and Vazacopoulos (1998) suggest several elaborate reoptimisation schemes as yet there is no strategy to indicate how this should be done. In addition no method is available to decide ....

....et al. Deterministic Job Shop Scheduling: Past, Present and Future 03 10 98 page(10) 1995) utilise an exact scheme to deal with DPCs. The most recent work (Balas and Vazacopoulos 1998) amalgamates guided variable search with SBP producing one of the best P J approaches available. Eventhough Balas and Vazacopoulos (1998) suggest several elaborate reoptimisation schemes as yet there is no strategy to indicate how this should be done. In addition no method is available to decide the size of the subproblem or which machine(s) to fix and in order to solve problems where job routings are more structured SBP will need ....

[Article contains additional citation context not shown here]

Balas, E., and Vazacopoulos, A., (1998), "Guided local search with shifting bottleneck for job-shop scheduling", Management Science 44/2, Feb, 262-275.

....does not guarantee a monotonic decrease of the makespan and the final SBI solution can be infeasible. As a result Dauz re P r s and Lasserre (1993) propose a heuristic strategy while Dauz reP r s (1995) and Balas et al. 1995) utilise an exact scheme to deal with DPCs. The most recent work (Balas and Vazacopoulos 1998) embeds a variable depth guided local search procedure (GL) within SBP. GL applies an interchange scheme based on a local neighbour hood structure that reverses more than one disjunction at a time. 3.2.1 COMPARATIVE ANALYSIS Table 7 compares the results of the various SBP methods applied to P J ....

....within SBP. GL applies an interchange scheme based on a local neighbour hood structure that reverses more than one disjunction at a time. 3.2.1 COMPARATIVE ANALYSIS Table 7 compares the results of the various SBP methods applied to P J . From table 7 it is clear that the strategy proposed by Balas and Vazacopoulos (1998) is superior, achieving the best solutions on all of the problems tested. This approach is also able to achieve the best known upper bound for the open problem ABZ 9. The primary weakness of this algorithm however is the high computing effort page (14) A state of the art review of job shop ....

[Article contains additional citation context not shown here]

Balas, E., and Vazacopoulos, A. (1998) Guided Local Search with Shifting Bottleneck for Job-Shop Scheduling, Management Science, Feb, 44(2), 262-275.

....RE lopt RE rand 3047 0.65 1.14 3045 0.66 1.16 HARD (10 Theta 50) 3173 0.66 1.12 2968 0.67 1.17 3022 0.68 1.15 2924 0.13 0.55 2794 0.18 0.58 EASY (10 Theta 50) 2852 0.11 0.53 2843 0.25 0.68 2823 0.12 0. 54 The best known solutions to the hard problems have been found by Balas and Vazacopoulos [2]. With the exception of one problem (makespan 2968) none of the hard problems is proofed to be solved to optimality while all easy problems are. For the ten instances two pools each containing 1 000 solutions are generated. One pool (rand) contains random solutions of S and the other pool (lopt) ....

E. Balas, A. Vazacopoulos, Guided local search with the shifting bottleneck for job shop scheduling, Report #MSRR-609, Graduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, PA 1994.

No context found.

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job-shop scheduling. Management Science, 44(2):262--275, 1998. 28

No context found.

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job-shop scheduling. Management Science, 44(2), 1998.

No context found.

E. Balas and A. Vazacopoulos. Guided local search with shifting bottleneck for job-shop scheduling. Management Science, 44(2), 1998.

No context found.

E. Balas, A. Vazacopoulos, Guided local search with shifting bottleneck for job-shop scheduling, Management Science 44 (2) (1998) 262--275.

No context found.

Balas, E. and Vazacopoulos, A. (1995). Guided local search with shifting bottleneck for job shop scheduling. Management Science Research Report, 609.

No context found.

E. Balas, A. Vazacopoulos, 1998. "Guided Local Search with Shifting Bottleneck for Job Shop Scheduling", Management Science 44, 262-275.

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