Restaurant Competition Optimization: A Novel Human-inspired Metaheuristic for Solving Real World Optimization Applications

In this paper, a novel metaheuristic algorithm, Restaurant Competition Optimization (RCO), is introduced, inspired by restaurants' competitive behavior to achieve a higher score than a food tester or critic.The main motivation of this research is to use a simple, intuitive, yet effective human behavior model to achieve an efficient balance between global search and local improvement in complex optimization problems.In RCO, each restaurant is modeled as a search agent, and the tester's scoring and ranking process is directly mapped to the objective function, so that competition, emulation of successful restaurants, and controlled innovations form the core of the search mechanism.The algorithm structure includes two complementary phases of exploration and exploitation, which cover, respectively, large changes and large movements in the solution space, as well as careful and gradual refinements around the best solutions.All human behaviors in the competition between restaurants are mapped to mathematical relationships in a reasonable, transparent manner, and there is no need to adjust complex external parameters.In order to evaluate the efficiency of RCO, this algorithm is implemented on four well-known engineering design problems from real-world applications, including the design of tension/compression springs, welded beams, speed reducers, and pressure vessels, and its performance is compared with nine well-known metaheuristic algorithms.Results from several independent runs show that RCO consistently provides more optimal values, less dispersion, and greater stability than competing algorithms.These findings indicate RCO's strong capability for solving nonlinear and constrained engineering optimization problems and its high potential for widespread application in real-world settings.