抄録
In this paper, interpolation and extrapolation search algorithms for the multi point SA were developed and the discussed in the continuous problems. In the multi point search, the next new searching point is determined from the information of the old searching points or is determined randomly. The former method is also classified into two categories; interpolation search and extrapolation search. To perform the effective search in the continuous problems, interpolation and extrapolation searches should be discussed carefully and should design good algorithms. This paper proposed the definition of the neighborhood of solutions. Then, the algorithms of interpolation and extrapolation searches were proposed. These two searches are dynamically changed. The effectiveness of the proposed methods was discussed through the numerical test function. From these results, it was described that the interpolation search increases the local search ability and the extrapolation search increases the global search ability.
