A Stability Analysis Based Parameter Setting Method for Spiral Optimization
Kenichi Tamura, Keiichiro Yasuda · 2013
In recent years, the authors proposed an effective metaheuristics method for continuous optimization problems based on analogy of spiral phenomena in nature which is called Spiral Optimization (SPO). The SPO has two setting parameters: the convergence rate and the rotation rate. Their values affect the search performance depending on computational and/or problem conditions. However, their effective setting methods without trial and error have not studied so far including analyses of its search dynamics which are needed for making such methods. This paper especially focuses on the convergence rate and proposes 1 its effective setting method from analyzing stability of dynamic equilibrium point of the SPO model. The effectiveness of the proposed method is confirmed thorough simulation for some benchmark functions and comparison with other representative metaheuristics methods.