Numerical optimization using synergetic swarms of foraging bacterial populations
| Autor: | Sotirios P. Chatzis, Spyros Koukas |
|---|---|
| Rok vydání: | 2011 |
| Předmět: |
Artificial intelligence
Computer and Information Sciences Mathematical optimization Expert systems (Computer science) Optimization problem Meta-optimization Computer science Foraging Population Swarming (honey bee) Local optimum Artificial Intelligence Derivative-free optimization Multi-swarm optimization education Metaheuristic education.field_of_study Bacteria General Engineering Population--Statistics Particle swarm optimization Bacteriology Computer Science Applications Global optimum Engineering and Technology Algorithms |
| Zdroj: | Expert Systems with Applications. |
| ISSN: | 0957-4174 |
| DOI: | 10.1016/j.eswa.2011.06.031 |
| Popis: | The bacterial foraging optimization (BFO) algorithm is a popular stochastic, population-based optimization technique that can be applied to a wide range of problems. Two are the major issues the BFO algorithm is confronted with: first, the foraging mechanism of BFO might in some cases induce the attraction of bacteria gathered near the global optimum by bacteria gathered to local optima, thus slowing down the whole population convergence. Second, BFO is susceptible to the curse-of-dimensionality, which makes it significantly harder to find the global optimum of a high-dimensional problem, compared to a low-dimensional problem with similar topology. In this paper, we introduce a novel BFO-based optimization algorithm aiming to address these issues, the synergetic bacterial swarming optimization (SBSO) algorithm. Our novel approach consists of: (i) the introduction of the swarming dynamics of the particle swarm optimization algorithm in the context of BFO, in order to ameliorate the convergence issues of the BFO bacteria foraging mechanism; and (ii) the utilization of multiple populations to optimize different components of the solution vector cooperatively, so as to mitigate the curse-of-dimensionality issues of the algorithm. We demonstrate the efficacy of our approach on several benchmark optimization problems. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect