Niche evolution strategy for global optimization

Many real-world problems can be formulated as global optimization problems. In recent years, evolutionary computation has been successfully applied in many such practical optimization problems which are hard to solve using traditional approaches due to their analytical intractability. However, practitioners are always in need of more effective and robust evolutionary algorithms as real-world problems become increasingly complex. In this paper, a niche evolution strategy (NES) motivated by S. Wright's (1931) shifting balance theory is therefore proposed. This NES employs several niches which evolve simultaneously and independently. During the evolutionary process, niche extinction and regeneration accompanied by gene flow occur at intervals of several generations. Two famous benchmark optimization problems are used to test the effectiveness of the proposed NES. The test results are compared with the corresponding results obtained from traditional single-population evolutionary computation and from distributed genetic algorithms (DGAs). It is shown that the NES is more effective than both of these approaches. It is also shown that the proposed NES can be readily applied in parallel computer architectures.