Iterated local search vs. hyper-heuristics: Towards general-purpose search algorithms

Burke, Edmund K.; Curtois, Timothy; Hyde, Matthew; Kendall, Graham; Ochoa, Gabriela; Petrović, Sanja; Vázquez-Rodríguez, José Antonio; Gendreau, Michel

doi:10.1109/cec.2010.5586064

Cited by 50 publications

(53 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Inspired from memetic algorithms [23,24], in which solutions are im-proved through successive application of mutation, crossover and hill climbing, the Robinhood hyper-heuristic uses the same ordering of groups and randomly fixing the ordering of heuristics within each group at a stage. There is also a strong empirical evidence in the literature that this ordering is a good choice even for selection hyper-heuristics as reported in [12,8]. Our hyper-heuristic uses the same ordering in the subsequent stage if there is an improvement in the solution quality at a given stage.…”

Section: Algorithm 1 Robinhood Hyper-heuristic Frameworkmentioning

confidence: 91%

“…Before CHeSC 2011 (Cross-Domain Heuristic Search Challenge), a mock competition was organised with hyflex and the performance of several well known previously proposed hyper-heuristics were compared across a subset of CHeSC problem domains. Burke et al [12] reported that the best performing hyper-heuristic was an iterated local search approach which applied a randomly selected mutational and ruin and re-create heuristic and then the hill climbers in a predefined sequence. This framework is based on the most successful hyper-heuristic framework reported to perform the best in [8].Özcan and Kheiri [13] provide a greedy heuristic selection strategy named dominancebased heuristic selection which aims to determine low level heuristics with good performance based on the trade-off between the change (improvement) in the solution quality and the number of steps taken.…”

Section: Related Workmentioning

confidence: 99%

“…A modified version of the adaptive acceptance method in [12] is introduced for the move acceptance. This acceptance method accepts all improvements as usual, but the deteriorations are accepted with respect to an adaptively changing rate, denoted as acceptanceRate.…”

Section: Algorithm 1 Robinhood Hyper-heuristic Frameworkmentioning

confidence: 99%

See 2 more Smart Citations

A Hyper-heuristic with a Round Robin Neighbourhood Selection

Kheiri

Özcan

2013

Evolutionary Computation in Combinatorial Optimization

View full text Add to dashboard Cite

Abstract. An iterative selection hyper-heuristic passes a solution through a heuristic selection process to decide on a heuristic to apply from a fixed set of low level heuristics and then a move acceptance process to accept or reject the newly created solution at each step. In this study, we introduce Robinhood hyper-heuristic whose heuristic selection component allocates equal share from the overall execution time for each low level heuristic, while the move acceptance component enables partial restarts when the search process stagnates. The proposed hyper-heuristic is implemented as an extension to a public software used for benchmarking of hyper-heuristics, namely HyFlex. The empirical results indicate that Robinhood hyper-heuristic is a simple, yet powerful and general multistage algorithm performing better than most of the previously proposed selection hyper-heuristics across six different Hyflex problem domains.

show abstract

Section: Algorithm 1 Robinhood Hyper-heuristic Frameworkmentioning

confidence: 91%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Hyper-heuristic with a Round Robin Neighbourhood Selection

Kheiri

Özcan

2013

Evolutionary Computation in Combinatorial Optimization

View full text Add to dashboard Cite

show abstract

“…-Iterated Local Search (ILS) [16]; few variations of this algorithm have been implemented. We implemented the ILS described in [2], an ILS that uses the VND variants described above as subsidiary local search, and a Hierarchical ILS [11]. Hierarchical ILS uses an ILS as subsidiary local search and applies a strong perturbation in the outer ILS and a small perturbation in the inner ILS.…”

Section: Design and Implementation Of Algorithmic Schematamentioning

confidence: 99%

A Non-adaptive Stochastic Local Search Algorithm for the CHeSC 2011 Competition

Mascia

Stützle

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In this work, we present our submission for the Cross-domain Heuristic Search Challenge 2011. We implemented a stochastic local search algorithm that consists of several algorithm schemata that have been offline-tuned on four sample problem domains. The schemata are based on all families of low-level heuristics available in the framework used in the competition with the exception of crossover heuristics. Our algorithm goes through an initial phase that filters dominated low-level heuristics, followed by an algorithm schemata selection implemented in a race. The winning schemata is run for the remaining computation time. Our algorithm ranked seventh in the competition results. In this paper, we present the results obtained after a more careful tuning, and a different combination of algorithm schemata included in the final algorithm design. This improved version would rank fourth in the competition.

show abstract

“…In this paper, a simple heuristic algorithm based on iterated local search (ILS) is proposed to solve the WSRP. ILS is one of the most conceptually simple and robust algorithms (Burke et al 2010). The essential idea of ILS is that when the local search is trapped at a local optimum, the ILS perturbs the previously visited local optimum instead of generating a new initial solution, and then restarts the local search from this modified solution (Lourenço et al 2003).…”

Section: Introductionmentioning

confidence: 99%

Iterated local search for workforce scheduling and routing problems

2017

View full text Add to dashboard Cite

The integration of scheduling workers to perform tasks with the traditional vehicle routing problem gives rise to the workforce scheduling and routing problems (WSRP). In the WSRP, a number of service technicians with different skills, and tasks at different locations with pre-defined time windows and skill requirements are given. It is required to find an assignment and ordering of technicians to tasks, where each task is performed within its time window by a technician with the required skill, for which the total cost of the routing is minimized. This paper describes an iterated local search (ILS) algorithm for the WSRP. The performance of the proposed algorithm is evaluated on benchmark instances against an off-the-shelf optimizer and an existing adaptive large neighbourhood search algorithm. The proposed ILS algorithm is also applied to solve the skill vehicle routing problem, which can be viewed as a special case of the WSRP. The computational results indicate that the proposed algorithm can produce high-quality solutions in short computation times.

show abstract

Iterated local search vs. hyper-heuristics: Towards general-purpose search algorithms

Cited by 50 publications

References 26 publications

A Hyper-heuristic with a Round Robin Neighbourhood Selection

A Hyper-heuristic with a Round Robin Neighbourhood Selection

A Non-adaptive Stochastic Local Search Algorithm for the CHeSC 2011 Competition

Iterated local search for workforce scheduling and routing problems

Contact Info

Product

Resources

About