ZyEvo: A Zygomycota-Inspired Dual-Phase Evolutionary Metaheuristic Algorithm

Two-dimensional search space optimization is important in many fields like engineering design, robotics path planning, and control systems. Conventional gradient-based techniques tend to perform poorly on difficult landscapes with multiple local minima or maxima. Population-based metaheuristics like Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and Differential Evolution (DE) have been found to perform well in such a scenario. Nonetheless, these techniques typically combine exploration and exploitation in one operator, such that they tend to be parameter-sensitive and prone to demanding tedious tuning.We introduce ZyEvo, a dual-phase metaheuristic inspired by binary fission and fusion, which explicitly separates intensification (elite cloning) and diversification (pair averaging). Augmented by uniform mutation and elitist selection, ZyEvo consistently outperforms seven state-of-the-art algorithms: Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Differential Evolution (DE), Bare Bones Fireworks Algorithm (BBFWA), Monarch Butterfly Optimization (MBO), and Harris Hawks Optimization (HHO).Across 2-D Rastrigin and Rosenbrock benchmarks, ZyEvo achieves up to 98% lower mean fitness and 35% faster convergence. Convergence test and Shannon-entropy diversity analyses demonstrate ZyEvo’s statistical and exploratory superiority.