Multi-population Fruit Fly Optimization Algorithm with Genetic Operators for Multi-target Path Planning

Automatic path planning is very important for many applications such as robots exploring unknown environments and logistics delivery. In this paper, we propose a discrete multi-population fruit fly optimization algorithm with genetic operators, where a greedy strategy is used to obtain good initial population, 3-opt heuristic search simulating olfactory to make the algorithm achieve higher convergence accuracy, multiple population collaborative strategy simulating vision to avoid the algorithm falling into local optima, and the genetic mechanism of selection-crossover-mutation prompts the population easily getting the optimal solution. The proposed planning algorithm has fewer parameters to be adjusted and has the advantages of high accuracy and fast convergence. The experimental results prove that the proposed planning algorithm performs best compared with other several algorithms.