Meta Morphic Particle Swarm Optimization

Particle Swarm Optimization is a simple and elegant optimization algorithm used to solve a large variety of different real-valued problems. When it comes to solving combinations of continuous and discrete problems however, PSO by itself is not very well suited for the task. There have been previous works addressing the issue of solving solely discrete problems with PSO, but solving problems involving both discrete and continuous parameters at the same time with a PSO-like algorithm has not yet been fully explored. In this paper we provide a novel PSO-based algorithm, called Meta Morphic Particle Swarm Optimization, which looks at solving a particular class of problems for which there exists a discrete set of possible ways to solve the problem where each possibility uses a different subset of a continuous, real-valued parameter space. We introduce a two-layered approach, a PSO in the inner layer for the continuous space, and an outer layer, guided migration scheme using probabilities to choose between the different possible solution sets. We analyze the performance and characteristics of this new algorithm and show how it can be used for real-world applications.


Published in:
Nature Inspired Cooperative Strategies for Optimization, 231-244
Presented at:
VI International Workshop on Nature Inspired Cooperative Strategies for Optimization, Canterbury, UK, September 2-4, 2013
Year:
2013
Publisher:
Springer
ISBN:
978-3-319-01691-7
Keywords:
Laboratories:


Note: The status of this file is: Involved Laboratories Only


 Record created 2014-02-16, last modified 2018-03-17

Preprint:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)