Effects of local search algorithms on groundwater remediation optimization using a self-adaptive hybrid genetic algorithm

Genetic algorithms allow solution of more complex, nonlinear civil, and environmental engineering problems than traditional gradient-based approaches, but they are more computationally intensive. One way to improve algorithm performance is through inclusion of local search, creating a hybrid genetic algorithm (HGA). The inclusion of local search helps to speed up the solution process and to make the solution technique more robust. This paper focuses on the effects of different local search algorithms on the performance of two different HGAs developed in previous phases of this research, the self-adaptive hybrid genetic algorithm (SAHGA) and the enhanced SAHGA. The algorithms are tested on eight test functions from the genetic and evolutionary computation literature and a groundwater remediation design case study. The results show that the selection of the local search algorithm to be combined with the simple genetic algorithm is critical to algorithm performance. The best local search algorithm varies for different problems, but can be selected prior to solving the problem by examining the reduction in fitness standard deviation associated with each local search algorithm, and the time distribution associated to the local search algorithm.