A Simple yet Efficient Function Optimizer
Xinsheng Lai · 2007
A new function optimization algorithm inspired by water flowing to the lowest point was proposed. We call it water flow optimization (WFO). No matter wherever the global lowest point locates could water find it if only enough water was poured into. We propose WFO by imitating the process of water flowing to the global lowest point. In this new algorithm, a population of parameter vector standing for a solution to the function optimization problem move to an attractive center which is continuously transferring. Verified by function optimization experiments, it is an efficient optimization algorithm. With an apt mutation probability, the WFO algorithm could efficiently avoid getting trapped by local optima, and it could find global optimum quickly and accurately as well. In addition, the algorithm has only two parameters: the population size and the mutation probability.