Heuristic generation via parameter tuning for online bin packing

Yarimcam, Ahmet; Asta, Shahriar; Özcan, Ender; Parkes, Andrew J.

doi:10.1109/eals.2014.7009510

Cited by 11 publications

(9 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A recent study byÖzcan and Parkes found that good (optimal) policies could actually be 'spiky' and complex [11,13]. Recent research in bin packing has tended to focus on metaheuristic strategies capable of automatically devising policies which are more complex than FF or BF and better suited to solving the problem [11,14,13].…”

Section: Previous Approaches For Online One-dimensional Bin Packingmentioning

confidence: 99%

Sparse, Continuous Policy Representations for Uniform Online Bin Packing via Regression of Interpolants

Drake

Swan

Neumann

et al. 2017

Evolutionary Computation in Combinatorial Optimization

View full text Add to dashboard Cite

Abstract. Online bin packing is a classic optimisation problem, widely tackled by heuristic methods. In addition to human-designed heuristic packing policies (e.g. first-or best-fit), there has been interest over the last decade in the automatic generation of policies. One of the main limitations of some previously-used policy representations is the trade-off between locality and granularity in the associated search space. In this article, we adopt an interpolation-based representation which has the jointlydesirable properties of being sparse and continuous (i.e. exhibits good genotype-to-phenotype locality). In contrast to previous approaches, the policy space is searchable via real-valued optimization methods. Packing policies using five different interpolation methods are comprehensively compared against a range of existing methods from the literature, and it is determined that the proposed method scales to larger instances than those in the literature.

show abstract

Section: Previous Approaches For Online One-dimensional Bin Packingmentioning

confidence: 99%

Sparse, Continuous Policy Representations for Uniform Online Bin Packing via Regression of Interpolants

Drake

Swan

Neumann

et al. 2017

Evolutionary Computation in Combinatorial Optimization

View full text Add to dashboard Cite

show abstract

“…Hence, online bin packing forms a good playground for testing the proposed approach. The first results (reported in [24] and later in [4] and [36]) indicates that the GA approach finds high quality policies for the specific packing problems that perform significantly better than the generic 'human designed' heuristics, such as, first fit and best fit [27,10]. In this paper, we show that by modifying mutation probabilities using tensor analysis, the performance of this framework can be improved significantly (when compared to the results achieved for our implementation of the GA approach in [24] as well as standard heuristics) on almost all instances.…”

Section: Introductionmentioning

confidence: 99%

“…In [5] an Apprenticeship Learning approach was proposed where an agent observes and learns the actions of high quality policy matrices during packing and generalizes its knowledge to unseen instances of various sizes exhibiting high performance particularly for larger instances. In [36], policy matrices are seen as heuristics with many parameters and are approached from a parameter tuning perspective. The Irace package was used to tune policies during training.…”

Section: Policy Matrix Representationmentioning

confidence: 99%

A Tensor Analysis Improved Genetic Algorithm for Online Bin Packing

Asta

Özcan

2015

Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

Mutation in a Genetic Algorithm is the key variation operator adjusting the genetic diversity in a population throughout the evolutionary process. Often, a fixed mutation probability is used to perturb the value of a gene. In this study, we describe a novel data science approach to adaptively generate the mutation probability for each locus. The trail of high quality candidate solutions obtained during the search process is represented as a 3 rd order tensor. Factorizing that tensor captures the common pattern between those solutions, identifying the degree of mutation which is likely to yield improvement at each locus. An online bin packing problem is used as an initial case study to investigate the proposed approach for generating locus dependent mutation probabilities. The empirical results show that the tensor approach improves the performance of a standard Genetic Algorithm on almost all classes of instances, significantly.

show abstract

“…Yarimcam et al (2014) showed that applying a parameter tuning approach does not match the performance of GA in the overall. Asta et al (2013b) showed that k-means clustering can be used to generalise the behaviour of GA for solving a given online bin packing problem.…”

Section: Introductionmentioning

confidence: 99%

“…The authors formulated the whole process as a special type of parameter tuning (Smit and Eiben, 2009;Yarimcam et al, 2014) in which the number of parameters is much larger than usually considered.…”

Section: Introductionmentioning

confidence: 99%

CHAMP: Creating heuristics via many parameters for online bin packing

Asta

Özcan

Parkes

2016

Expert Systems with Applications

Self Cite

View full text Add to dashboard Cite

The online bin packing problem is a well-known bin packing variant which requires immediate decisions to be made for the placement of a lengthy sequence of arriving items of various sizes one at a time into fixed capacity bins without any overflow. The overall goal is maximising the average bin fullness. We investigate a 'policy matrix' representation which assigns a score for each decision option independently and the option with the highest value is chosen for one dimensional online bin packing. A policy matrix might also be considered as a heuristic with many parameters, where each parameter value is a score.We hence investigate a framework which can be used for creating heuristics via many parameters. The proposed framework combines a Genetic Algorithm optimiser, which searches the space of heuristics in policy matrix form, and an online bin packing simulator, which acts as the evaluation function. The empirical results indicate the success of the proposed approach, providing the best solutions for almost all item sequence generators used during the experiments.We also present a novel fitness landscape analysis on the search space of policies. This study hence gives evidence of the potential for automated discovery by intelligent systems of powerful heuristics for online problems; reducing the need for expensive use of human expertise.

show abstract

Heuristic generation via parameter tuning for online bin packing

Cited by 11 publications

References 30 publications

Sparse, Continuous Policy Representations for Uniform Online Bin Packing via Regression of Interpolants

Sparse, Continuous Policy Representations for Uniform Online Bin Packing via Regression of Interpolants

A Tensor Analysis Improved Genetic Algorithm for Online Bin Packing

CHAMP: Creating heuristics via many parameters for online bin packing

Contact Info

Product

Resources

About