This paper provides a novel meta-heuristic optimization algorithm for solving continuous optimization problems efficiently in the field of numerical and engineering optimization: Piranha Foraging Optimization Algorithm (PFOA). The algorithm is inspired by the flexible and mobile foraging behaviour of piranha swarm and divides their foraging behavior into three patterns: localized group attack, bloodthirsty cluster attack and scavenging foraging, simulates the above behaviors to construct two dynamic search processes for exploration and exploitation. PFOA uses three strategies of non-linear parameter control, population survival and reverse evasion search to enable populations to have better population diversity at different stages of the search and to help find better solutions. To gain insight into the performance of PFOA, visualization methods were used to assess the efficiency of PFOA optimization and to analyse the impact of the characteristics of the three foraging modes, the sensitivity of the parameters and the size of the piranha population on the algorithm. The algorithm performance was further tested with 27 CEC benchmark functions and four real engineering design optimization problems, and the results were compared with 13 well-known meta-heuristics. Test results based on statistical methods such as box-line plots, Wilcoxon rank sum test and Friedman test in multiple dimensions (30, 50, 100 and fixed dimensions) show significant differences compared to other algorithms and that the performance of the algorithm is stable and in significant improvement. The unique advantages of PFOA in terms of the equilibrium of convergence speed and exploration can avoid getting trapped in local optimum regions and effectively solve optimization problems in complex search spaces.

No abstract