## Directing the Search of Evolutionary and Neighbourhood-Search Optimisers for the Flowshop Sequencing Problem with an Idle-Time Heuristic (1997)

Venue: | In The 1997 AISB Workshop on Evolutionary Computing |

Citations: | 2 - 1 self |

### BibTeX

@INPROCEEDINGS{Ross97directingthe,

author = {Peter Ross and Andrew Tuson},

title = {Directing the Search of Evolutionary and Neighbourhood-Search Optimisers for the Flowshop Sequencing Problem with an Idle-Time Heuristic},

booktitle = {In The 1997 AISB Workshop on Evolutionary Computing},

year = {1997}

}

### OpenURL

### Abstract

. This paper presents a heuristic for directing the neighbourhood (mutation operator) of stochastic optimisers, such as evolutionary algorithms, so to improve performance for the flowshop sequencing problem. Based on idle time, the heuristic works on the assumption that jobs that have to wait a relatively long time between machines are in an unsuitable position in the schedule and should be moved. The results presented here show that the heuristic improves performance, especially for problems with a large number of jobs. In addition the effectiveness of the heuristic and search in general was found to depend upon the neighbourhood structure in a consistent fashion across optimisers. 1 Introduction One problem of interest in the artificial intelligence and operations research scheduling communities is the flowshop sequencing problem. The aim of this problem is, quite simply, to find a sequence of n jobs that minimises the makespan --- the time for all of the jobs to be proces...

### Citations

3529 | Optimization by simulated annealing
- Gelatt, Vecchi
- 1983
(Show Context)
Citation Context ...istribution. 3 The Optimisation Techniques Considered A variety of stochastic optimisation methods (an overview is given in [4]) were examined: Stochastic Hillclimbing (SHC), Simulated Annealing (SA) =-=[8]-=-, Threshold Accepting (TA) [3], Record-to-Record Travel (RTRT) [2], and an Evolutionary Algorithm (EA) [6]. Basic implementations were used in this study as the aim was to investigate whether the neig... |

815 |
Adaptive in Natural and Artificial Systems. Ann Arbor
- Holland
- 1975
(Show Context)
Citation Context ...rview is given in [4]) were examined: Stochastic Hillclimbing (SHC), Simulated Annealing (SA) [8], Threshold Accepting (TA) [3], Record-to-Record Travel (RTRT) [2], and an Evolutionary Algorithm (EA) =-=[6]-=-. Basic implementations were used in this study as the aim was to investigate whether the neighbourhood could be usefully directed, and not to find the most effective optimiser. The encoding used for ... |

152 | Benchmarks of basic scheduling problems
- Taillard
- 1993
(Show Context)
Citation Context ...J i ; j) = maxfC(J i\Gamma1 ; j); C(J i ; j \Gamma 1)g + p(J i ; j) for i = 2; :::; n; j = 2; :::; m Cmax = C(Jn ; m) Standard benchmarks exist for this problem, and were used in this study. Taillard =-=[16]-=- produced a set of test problems which, using a very lengthy Tabu Search procedure, were still inferior to their lower bounds, with problem sizes ranging from 20 jobs and 5 machines, to 500 jobs and 2... |

111 |
Threshold accepting: A general purpose optimization algorithm appearing superior to simulated annealing
- Dueck, Scheuer
- 1990
(Show Context)
Citation Context ...n Techniques Considered A variety of stochastic optimisation methods (an overview is given in [4]) were examined: Stochastic Hillclimbing (SHC), Simulated Annealing (SA) [8], Threshold Accepting (TA) =-=[3]-=-, Record-to-Record Travel (RTRT) [2], and an Evolutionary Algorithm (EA) [6]. Basic implementations were used in this study as the aim was to investigate whether the neighbourhood could be usefully di... |

99 |
Scheduling and Rescheduling with Iterative Repair
- Zweben, Daun, et al.
- 1994
(Show Context)
Citation Context ...irected mutation' operator [15] that has been successfully used in timetabling, where exams with a high number of constraint violations were more likely to be moved; it is also related to the work by =-=[18,19]-=- on iterative repair scheduling, and by [9] on the `min-conflicts' heuristic. For this problem, such a belief can be formulated by using the idle times, the time a job spends waiting between machines,... |

36 |
Simulated annealing for permutation flow-shop scheduling
- H, Potts
- 1989
(Show Context)
Citation Context ...200x20. The value of N used was solely dependent upon n and was set to 5000, 6750, 8000, and 9250 evaluations respectively. This is a roughly ln(n) relationship, justified from empirical results from =-=[12]-=-. 5.1 Stochastic Hillclimbing Results Problem Undirected Heuristic + Random Heuristic + Heuristic 20x5 1296.26 (3.63) 1291.68 (7.52) [T9] 1294.58 (5.68) [M4] 20x10 1633.32 (21.73) 1629.32 (21.15) [M5]... |

29 |
Genetic Algorithms for Function Optimization
- Brindle
- 1981
(Show Context)
Citation Context ...en the neighbourhood operator was then applied, the position(s) of the move operator was decided on the basis of the idle time of the job at that position, by using one of either Tournament Selection =-=[1]-=- or Marriage Selection. Marriage Selection is a variant of tournament selection designed to soften its inherently high selection pressure. Possible jobs are picked at random until either one is found ... |

26 | Improving evolutionary timetabling with delta evaluation and directed mutation
- Ross, Corne, et al.
- 1994
(Show Context)
Citation Context ...e the solution. These beliefs can then be captured in a heuristic, which makes it more likely that such moves are attempted first. The approach used here was based on the `directed mutation' operator =-=[15]-=- that has been successfully used in timetabling, where exams with a high number of constraint violations were more likely to be moved; it is also related to the work by [18,19] on iterative repair sch... |

23 | Scheduling by Genetic Local Search with Multi-Step Crossover
- Yamada, Ryohei
- 1996
(Show Context)
Citation Context ...n a constructive algorithm. Also, the concept of a `candidate list', which orders the sequence in which the moves are considered has been used by [11], and in the evolutionary computing literature by =-=[17]-=-. However, the authors are unaware of an idle-time heuristic being used for evolutionary and stochastic neighbourhood search optimisation approaches, in this way, for this problem and whether, or to w... |

21 |
A fast tabu search algorithm for the permutation flowshop problem
- Nowicki, Smutnicki
- 1996
(Show Context)
Citation Context ... deterministic hillclimber, and [13] used idle times in a constructive algorithm. Also, the concept of a `candidate list', which orders the sequence in which the moves are considered has been used by =-=[11]-=-, and in the evolutionary computing literature by [17]. However, the authors are unaware of an idle-time heuristic being used for evolutionary and stochastic neighbourhood search optimisation approach... |

13 |
A new heuristic for the n-job, m-machine flow shop problem
- Ho, Chang
- 1991
(Show Context)
Citation Context ...ochastic optimisation techniques will be examined. It should be noted that the idea of using idle times to improve the performance of heuristic methods for the flowshop sequencing problem is not new: =-=[5]-=- proposed the use of a similar heuristic in an approach that is essentially a deterministic hillclimber, and [13] used idle times in a constructive algorithm. Also, the concept of a `candidate list', ... |

10 |
The Vehicle Routing Problem and Local Search Metaheuristics
- Hjorring
- 1995
(Show Context)
Citation Context ... problem --- all processing times were generated randomly from a U(1; 100) distribution. 3 The Optimisation Techniques Considered A variety of stochastic optimisation methods (an overview is given in =-=[4]-=-) were examined: Stochastic Hillclimbing (SHC), Simulated Annealing (SA) [8], Threshold Accepting (TA) [3], Record-to-Record Travel (RTRT) [2], and an Evolutionary Algorithm (EA) [6]. Basic implementa... |

9 |
Optimising flowshop scheduling through adaptive genetic algorithms
- Mott
- 1990
(Show Context)
Citation Context ... 2 8 5 4 2 7 8 1 6 4 3 5 4 2 2 8 3 6 7 1 5 7 8 Legalise 1 6 3 4 3 5 2 7 8 6 7 5 2 8 Crossover 1 4 Fig. 4. The Modified-PMX Crossover Operator The crossover operator used for the EA was `Modified PMX' =-=[10]-=-, found to give good results on this problem. This operator performs a two-point crossover upon the two strings selected. The repair procedure then analyses one string for duplicates: when one is foun... |

7 |
New Optimisation Heuristics: The Great Deluge Algorithm and the Record-to-Record Travel
- Dueck
- 1990
(Show Context)
Citation Context ... stochastic optimisation methods (an overview is given in [4]) were examined: Stochastic Hillclimbing (SHC), Simulated Annealing (SA) [8], Threshold Accepting (TA) [3], Record-to-Record Travel (RTRT) =-=[2]-=-, and an Evolutionary Algorithm (EA) [6]. Basic implementations were used in this study as the aim was to investigate whether the neighbourhood could be usefully directed, and not to find the most eff... |

5 |
An efficient heuristic approach to the scheduling of jobs in a flowshop
- Rajendran, Chaudhuri
- 1991
(Show Context)
Citation Context ...ve the performance of heuristic methods for the flowshop sequencing problem is not new: [5] proposed the use of a similar heuristic in an approach that is essentially a deterministic hillclimber, and =-=[13]-=- used idle times in a constructive algorithm. Also, the concept of a `candidate list', which orders the sequence in which the moves are considered has been used by [11], and in the evolutionary comput... |

4 |
Solving Large Scale CSP and Scheduling Problems with a Heuristic Repair Method
- Minton, Philipps, et al.
- 1990
(Show Context)
Citation Context ...uccessfully used in timetabling, where exams with a high number of constraint violations were more likely to be moved; it is also related to the work by [18,19] on iterative repair scheduling, and by =-=[9]-=- on the `min-conflicts' heuristic. For this problem, such a belief can be formulated by using the idle times, the time a job spends waiting between machines, associated with each job in the sequence t... |

1 |
Rinnooy Kan. Machine Sequencing Problems: Classification, complexity and computations. Martinus Nijhoff, The Hague
- G
- 1976
(Show Context)
Citation Context ...quences can lead to temporary blockages in the flowline because completed stages cannot proceed, as a stage ahead in the flowline is still being processed (Figure 1). This task is known to be NP-hard =-=[7]-=- (the number of possible sequences is n!) and can be formalised as follows: n jobs have to be processed (in the same order) on m machines; the aim is to find a job permutation fJ 1 ; J 2 ; :::; Jn g s... |

1 |
Learning to Improve Iterative Repair Scheduling
- Zweben, Davis
- 1992
(Show Context)
Citation Context ...irected mutation' operator [15] that has been successfully used in timetabling, where exams with a high number of constraint violations were more likely to be moved; it is also related to the work by =-=[18,19]-=- on iterative repair scheduling, and by [9] on the `min-conflicts' heuristic. For this problem, such a belief can be formulated by using the idle times, the time a job spends waiting between machines,... |