Effectiveness research of a global optimization algorithm inspired by some aspects of cockroach behaviour

Consider the Roach Infestation Optimization (RIO) algorithm, which belongs to the class of population-based algorithms inspired by wildlife. The RIO algorithm was proposed in 2008 and can be considered a deep modification of the well-known and one of the most effective particle swarm optimization (PSO) algorithms. The interest in the RIO algorithm is due to the fact that, due to the high efficiency of the PSO algorithm for a wide range of global optimization problems, the study of the modification of this algorithm, which is represented by the RIO algorithm, is of particular interest. The purpose of the work is to implement software and study the efficiency of the RIO algorithm for the well-known complex multimodal test functions of Ratrigin and Ackley. A feature of the study is the search for a global extremum (minimum) of these functions in a wide region of the search space, in which the number of local minima of these functions is extremely large. We present the formulation of the considered global optimization problem, as well as a description of the RIO algorithm, a distinctive feature of which is the use not of the original designations of the authors of this algorithm, but of the unified designations that we use when considering other population algorithms. We describe the software that implements the algorithm and the organization of computational experiments to study its effectiveness. Finally, the article presents the research results showing the high prospects of the RIO algorithm for solving global optimization problems.

1. 1. Karpenko A.P. Sovremennye algoritmy poiskovoj optimizacii. Algoritmy vdohnovlennye prirodoj. Moskva: Izdatel'stvo MGTU im. N.E. Baumana; 2014.

2. 2. Havens T.C. et al. Roach infestation optimization. In: Proceedings of the 2008 IEEE Swarm Intelligence Symposium, St. Louis, MO, USA. 2008:21–23.

3. 3. Bo Xing, Wen-Jing Gao. Innovative Computational Intelligence: A Rough Guide to 134 Clever Algorithms. Springer International Publishing Switzerland; 2014.

4. 4. Jeanson R. et al. Selforganized aggregation in cockroaches. Animal Behaviour. 2005;69:169–180.

5. 5. Halloy J. et al. Social integration of robots into groups of cockroaches to control self-organizined choices. Science. 2007; 318(5853):1155-1158.

6. 6. Ame J. at al. Collegial decision making based on social amplification leads to optimal group formation. Proc. Natl. Acad. Sci. 2006;103(15):5835–5840.

7. 7. Garnier S. et al. Collective decision-making by a group of cockroach-like robots. Proc. 2005 IEEE Swarm Intelligence Symposium (SIS 2005). Pasadena, CA, USA. 2005;233-240.

8. 8. Watanabe H., Mizunami M. Pavolv’s cockroach: Classical conditioning of salivation in an insect. PLoS ONE. 2007;2(6):529.

9. 9. Kennedy J., Eberhardt R. Particle swarm optimization. Proceedings of the IEEE Int. Conf. on Neural Networks, Piscataway, NJ. 1995;1942–1948.

10. 10. Clerc M. Particle Swarm Optimization. Newport Beach, CA: ISTE USA, 2006.