RANDOM NUMBER GENERATOR INFLUENCE ON DIFFERENTIAL SEARCH PERFORMANCE

摘要

Differential Search is an optimization approach that simulates the Brownian like random walk movement of migration organisms. Its inner workings relay on Random Number Generators (RNG) to emulate the probability of the natural driven process and therefore, the properties and performance of the RNG used has an influence on the overall algorithm performance. This study focuses on assessing this influence and on the manner in which the best suited RNG must be selected in order to achieve the optimal results for different types of problems. The results showed that depending on the properties of the problem, different RNGs provide the best solutions. However, as the number of function evaluation rises in report to the problem dimensionality, the influence on the RNG becomes smaller.