A Hierarchical Particle Swarm Optimizer for Dynamic Optimization Problems

Abstract

Abstract. Particle Swarm Optimization (PSO) methods for dynamic function optimization are studied in this paper. We compare dynamic variants of standard PSO and Hierarchical PSO (H-PSO) on different dynamic benchmark functions. Moreover, a new type of hierarchical PSO, called Partitioned H-PSO (PH-PSO), is proposed. In this algorithm the hierarchy is partitioned into several sub-swarms for a limited number of generations after a change occurred. Different methods for determining the time when to rejoin the hierarchy and how to handle the topmost sub-swarm are discussed. The test results show that H-PSO performs significantly better than PSO on all test functions and that the PH-PSO algorithms often perform best on multimodal functions where changes are not too severe.

A survey of swarm intelligence for dynamic optimization: Algorithms and applications

“…Janson and Middendorf [162] proposed a partitioned hierarchical PSO (PH-PSO) that uses a dynamic tree-based neighbourhood structure. Each particle is placed on a single node of the tree and particles are arranged hierarchically according to their fitness.…”

A new discrete particle swarm algorithm applied to attribute selection in a bioinformatics data set

“…Most of the particle swarm algorithms present in the literature deal only with continuous variables [1,5,10]. This is a significant limitation because many optimization problems are set in a space featuring discrete variables.…”

Dynamic Problems and Nature Inspired Meta-Heuristics

“…This gives a dynamic neighbourhood that does require extensive calculation. This has been adapted to dynamic problems by Janson and Middendorf [23,24]. After the value of the best-known position (gbest) changes (it is reevaluated every cycle) a few sub-swarms are reinitialised while the rest are reset (have their old personal best information erased and replaced with the current position).…”

“…For dynamic problems this tabu list can be completely removed on the grounds that any particular point in problem space may be a good optimum at several disjoined times. Alternatively extending the idea from Janson and Middendorf [24], each of these previously explored optima could be periodically re-examined and only those points whose fitness had significantly changed are removed from the tabu lists. It is not clear at this stage how the evolutionary pressure that is an important part of WoSP would respond to a dynamic problem.…”