Jacomine Grobler

Alternative hyper-heuristic strategies for multi-method global optimization

The purpose of this paper is to investigate the use of meta-heuristics as low-level heuristics in a hyper-heuristic framework. A novel multi-method hyper-heuristic algorithm which makes use of a number of common meta-heuristics is presented. Algorithm performance is evaluated on a diverse set of real parameter benchmark problems and meaningful conclusions are drawn with respect to the selection of alternative low-level heuristics and the acceptance of the obtained solutions within the proposed multi-method meta-heuristic approach.

Investigating the impact of alternative evolutionary selection strategies on multi-method global optimization

Algorithm selection is an important consideration in multi-method global optimization. This paper investigates the use of various algorithm selection strategies derived from well known evolutionary selection mechanisms. Selection strategy performance is evaluated on a diverse set of floating point benchmark problems and meaningful conclusions are drawn with regard to the impact of selective pressure on algorithm selection in a multi-method environment.

Heuristic space diversity control for improved meta-hyper-heuristic performance

This paper expands on the concept of heuristic space diversity and investigates various strategies for the management of heuristic space diversity within the context of a meta-hyper-heuristic algorithm in search of greater performance benefits. Evaluation of various strategies on a diverse set of floating-point benchmark problems shows that heuristic space diversity has a significant impact on hyper-heuristic performance. An exponentially increasing strategy (EIHH) obtained the best results. The value of a priori information about constituent algorithm performance on the benchmark set in question was also evaluated. Finally, EIHH demonstrated good performance when compared to a popular population based algorithm portfolio algorithm and the best performing constituent algorithm.

Hybridizing PSO and DE for improved vector evaluated multi-objective optimization

This paper introduces a new vector evaluated multi-objective optimization algorithm. The vector evaluated differential evolution particle swarm optimization (VEDEPSO) algorithm is a hybridization of the classical vector evaluated particle swarm optimization (VEPSO) and vector evaluated differential evolution (VEDE) algorithms of Parsopoulos et. al. [9], [10]. Comparisons of VEDEPSO with respect to VEPSO and VEDE on a well known multi-objective benchmark problem set indicated that significant performance improvements can be attributed to the VEDEPSO algorithm.

Multi-method algorithms: Investigating the entity-to-algorithm allocation problem

This paper investigates the algorithm selection problem, otherwise referred to as the entity-to-algorithm allocation problem, within the context of three recent multi-method algorithm frameworks. A population-based algorithm portfolio, a meta-hyper-heuristic and a bandit based operator selection method are evaluated under similar conditions on a diverse set of floating-point benchmark problems. The meta-hyper heuristic is shown to outperform the other two algorithms.

Investigating the use of local search for improving meta-hyper-heuristic performance

This paper investigates the use of local search strategies to improve the performance of a meta-hyper-heuristic algorithm, a hyper-heuristic which employs one or more meta-heuristics as low-level heuristics. Alternative mechanisms for selecting the solutions to be refined further by means of local search, as well as the intensity of subsequent refinement in terms of number of allowable function evaluations, are investigated. Furthermore, defining a local search as one of the low-level heuristics versus applying the algorithm directly to the solution space is also investigated. Performance is evaluated on a diverse set of floating-point benchmark problems. The addition of local search was found to improve algorithm results significantly. Random selection of solutions for further refinement was identified as the best selection strategy and a higher intensity of refinement was identified as most desirable. Better results were obtained by applying the local search algorithm directly to the search space instead of defining it as a low-level heuristic.

Multi-objective DE and PSO strategies for production scheduling

This paper investigates the application of alternative multi-objective optimization (MOO) strategies to a complex scheduling problem. Two vector evaluated algorithms, namely the vector evaluated particle swarm optimization (VEPSO) algorithm as well as the vector evaluated differential evolution (VEDE) algorithm is compared to a differential evolution based modified goal programming approach. This paper is considered significant since no other reference to the application of vector evaluated algorithms in a scheduling environment could be found. Algorithm performance is evaluated on real customer data and meaningful conclusions are drawn with respect to the application of MOO algorithms in a multiple machine multi-objective scheduling environment.

The entity-to-algorithm allocation problem: extending the analysis

This paper extends the investigation into the algorithm selection problem in hyper-heuristics, otherwise referred to as the entity-to-algorithm allocation problem, introduced by Grobler et al.. Two newly developed population-based portfolio algorithms (the evolutionary algorithm based on selfadaptive learning population search techniques (EEA-SLPS) and the Multi-EA algorithm) are compared to two metahyper- heuristic algorithms. The algorithms are evaluated under similar conditions and the same set of constituent algorithms on a diverse set of floating-point benchmark problems. One of the meta-hyper-heuristics are shown to outperform the other algorithms, with EEA-SLPS coming in a close second.

Solution Space Diversity Management in a Meta-hyperheuristic Framework

This paper investigates various strategies for the management of solution space diversity within the context of a meta-hyper heuristic algorithm. The adaptive local search meta-hyper heuristic (ALSHH), which adaptively applies a local search algorithm when the population diversity strays outside a predetermined solution space diversity profile, is proposed. ALSHH was shown to compare favourably with algorithms making use of local search and diversity maintenance strategies applied at constant intervals throughout the optimization run. Good performance is also demonstrated with respect to two other popular multi-method algorithms.

Headless Chicken Particle Swarm Optimization Algorithms

This paper investigates various strategies for implementing the headless chicken macromutation operator in the particle swarm optimization domain. Three different headless chicken particle swarm optimization algorithms are proposed and evaluated against a standard guaranteed convergence PSO algorithm on a diverse set of benchmark problems. Competitive performance is demonstrated by a Von Neumann headless chicken particle swarm optimization algorithm when compared to a classic guaranteed convergence particle swarm optimization algorithm. Statistically significantly superior results are obtained over a number of difficult benchmark problems.