Simulated annealing with search mechanisms of interpolation and extrapolatolation domain
Yuta Tomatsu, Masahiro Nakao, Tomoyuki Hiroyasu, Hisatake Yokouchi, Mitsunori Miki, Masato Yoshimi · World Automation Congress · 2010
Appropriate interpolation and extrapolation search methods are important in evolutionary computation. This paper presents a discussion of interpolation and extrapolation search methods in continuous optimization problems. First, we defined interpolation and extrapolation searches determined from the search domains, and introduce the stochastic multiple point search algorithm, which is based on simulated annealing, as an interpolation and extrapolation search algorithm. The effectiveness of interpolation and extrapolation search are discussed through numerical test functions. The results showed that interpolation search increases accuracy of the search for optimum solutions and extrapolation search increases the discovery rate of optimum solutions.