Neighborhood research approach in swarm intelligence for solving the optimization problems

Abstract

The article discusses the key problem of swarm algorithms and the bioinspired approach, which is to determine the proximity function and study the emerging neighborhoods in order to solve optimization problems. There is a detailed discussion of one of the most important design phases, namely, the VLSI components placement problem, whose solutions fineness directly affects the quality of circuit tracing. The solution of the neighborhoods and solution proximity problem is demonstrated by the study of the solutions by means of hybrid search methods. The main idea of this approach is the sequential use of genetic and swarm algorithms. We propose a new formation principle of the positions' neighborhood in the solution space based on the bee colony algorithm, which uses the concept of neighborhood in a circular search space. There are also experimental studies which show that the time complexity of the developed approach does not go beyond polynomial dependence.