An Intelligent Hyper-Heuristic Framework for CHeSC 2011

2016 IEEE Congress on Evolutionary Computation (CEC)

Atkin

2016

Abstract-Memetic algorithms, which hybridise evolutionary algorithms with local search, are well-known metaheuristics for solving combinatorial optimisation problems. A common issue with the application of a memetic algorithm is determining the best initial setting for the algorithmic parameters, but these can greatly influence its overall performance. Unlike traditional studies where parameters are tuned for a particular problem domain, in this study we do tuning that is applicable to cross-domain search. We extend previous work by tuning the parameters of a steady state memetic algorithm via a 'design of experiments' approach and provide surprising empirical results across nine problem domains, using a cross-domain heuristic search tool, namely HyFlex. The parameter tuning results show that tuning has value for cross-domain search. As a side gain, the results suggest that the crossover operators should not be used and, more interestingly, that single point based search should be preferred over a population based search, turning the overall approach into an iterated local search algorithm. The use of the improved parameter settings greatly enhanced the crossdomain performance of the algorithm, converting it from a poor performer in previous work to one of the stronger competitors.

Section: Hyper-heuristics Flexible Framework (Hyflex)mentioning

confidence: 99%

An investigation of tuning a memetic algorithm for cross-domain search

Gumus

2016 IEEE Congress on Evolutionary Computation (CEC)

Atkin

2016

“…The performance of the proposed hyper-heuristic is then compared to that of the two building block algorithms, namely SR-NA and SR-IE. Also, the current state-ofthe-art algorithm, AdapHH [41] is included in the comparisons. The interesting aspect of TeBHA-HH is that, generally speaking, it uses a hyper-heuristic based on random heuristic selection, decomposes the low level heuristics into two subsets and again applies the same hyper-heuristic using two simple move acceptance methods.…”

Section: Experiments On the Chesc 2011 Domainsmentioning

confidence: 99%

A tensor-based selection hyper-heuristic for cross-domain heuristic search

Asta

2015

Information Sciences

Hyper-heuristics have emerged as automated high level search methodologies that manage a set of low level heuristics for solving computationally hard problems. A generic selection hyper-heuristic combines heuristic selection and move acceptance methods under an iterative single point-based search framework. At each step, the solution in hand is modified after applying a selected heuristic and a decision is made whether the new solution is accepted or not. In this study, we represent the trail of a hyper-heuristic as a third order tensor. Factorization of such a tensor reveals the latent relationships between the low level heuristics and the hyper-heuristic itself. The proposed learning approach partitions the set of low level heuristics into two equal subsets where heuristics in each subset are associated with a separate move acceptance method. Then a multi-stage hyper-heuristic is formed and while solving a given problem instance, heuristics are allowed to operate only in conjunction with the associated acceptance method at each stage. To the best of our knowledge, this is the first time tensor analysis of the space of heuristics is used as a data science approach to improve the performance of a hyper-heuristic in the prescribed manner. The empirical results across six different problem domains from a benchmark indeed indicate the success of the proposed approach.

“…Higher values indicate that hill climbing approach searches more neighbourhoods for improvement. The top three selection hyper-heuristics that generalize well across the CHeSC 2011 problem domains are AdapHH [26], VNS-TW [27] and ML [28].…”

Section: Hyflex and First Cross-domain Heuristic Search Challengmentioning

confidence: 99%

An apprenticeship learning hyper-heuristic for vehicle routing in HyFlex

Asta

2014 IEEE Symposium on Evolving and Autonomous Learning Systems (EALS)

2014

Abstract-Apprenticeship learning occurs via observations while an expert is in action. A hyper-heuristic is a search method or a learning mechanism that controls a set of low level heuristics or combines different heuristic components to generate heuristics for solving a given computationally hard problem. In this study, we investigate into a novel apprenticeshiplearning-based approach which is used to automatically generate a hyper-heuristic for vehicle routing. This approach itself can be considered as a hyper-heuristic which operates in a train and test fashion. A state-of-the-art hyper-heuristic is chosen as an expert which is the winner of a previous hyper-heuristic competition. Trained on small vehicle routing instances, the learning approach yields various classifiers, each capturing different actions that the expert hyper-heuristic performs during the search process. Those classifiers are then used to produce a hyper-heuristic which is potentially capable of generalizing the actions of the expert hyperheuristic while solving the unseen instances. The experimental results on vehicle routing using the Hyper-heuristic Flexible (HyFlex) framework shows that the apprenticeship-learningbased hyper-heuristic delivers an outstanding performance when compared to the expert and some other previously proposed hyper-heuristics.