Usually, evolutionary algorithms keep the size of the population fixed. In the context of dynamic environments, many approaches divide the main population into two, one part that evolves as usual another that plays the role of memory of past good solutions. The size of these two populations is often chosen off-line. Usually memory size is chosen as a small percentage of population size, but this decision can be a strong weakness in algorithms dealing with dynamic environments. In this work we do an experimental study about the importance of this parameter for the algorithm's performance. Results show that tuning the population and memory sizes is not an easy task and the impact of that choice on the algorithm's performance is significant. Using an algorithm that dynamically adjusts the population and memory sizes outperforms standard approach.
展开▼
