摘要：Non–stationary, or dynamic, problems change overtime. There exist a variety of forms of dynamism.The concept of dynamic environments in thecontext of this paper means that the fitnesslandscape changes during the run of an evolutionaryalgorithm. Genetic diversity is crucial to provide thenecessary adaptability of the algorithm to changes.Two mechanism of macromutation are incorporatedto the algorithm to maintain genetic diversity in thepopulation. The algorithm was tested on a set ofdynamic testing functions provided by a dynamicfitness problem generator. The main goal was todeterminate the algorithm's ability to reacting tochanges of optimum values that alter their locations,so that the optimum value can still be tracked whendimensional and multimodal scalability in thefunctions is adjusted. The effectiveness andlimitations of the proposed algorithm is discussedfrom results empirically obtained
