Monkey search: a novel metaheuristic search for global optimization

TLDR

The authors introduce Monkey Search, a novel metaheuristic for global optimization inspired by monkeys climbing trees in search of food. Monkey Search represents tree branches as perturbations between neighboring feasible solutions, marks and updates promising branches during climbs, and allows a wide range of perturbations drawn from other metaheuristics. Experiments on Lennard‑Jones and Morse clusters and on protein folding models demonstrate that Monkey Search performs competitively with existing metaheuristics.

Abstract

We propose a novel metaheuristic search for global optimization inspired by the behavior of a monkey climbing trees looking for food. The tree branches are represented as perturbations between two neighboring feasible solutions of the considered global optimization problem. The monkey mark and update these branches leading to good solutions as it climbs up and down the tree. A wide selection of perturbations can be applied based on other metaheuristic methods for global optimization. We show that Monkey Search is competitive compared to the other metaheuristic methods for optimizing Lennard‐Jones and Morse clusters, and for simulating protein molecules based on a geometric model for protein folding.