The Human-Inspired Algorithm: A Hybrid Nature-Inspired Approach to Optimizing Continuous Functions with Constraints

摘要

The novel Human-Inspired Algorithm (HIA) uses a searching strategy based on people's intelligence to effectively find maximum or near maximum values of a continuous function with constraints. Initially, HIA evenly distributes points in equal subspaces of the whole search space and finds an elite subspace with the largest sum of function values. HIA uses the Genetic Algorithm (GA) to generate multiple sites from the elite subspace and the whole space and creates a hypercube with the best site as its center. HIA iteratively searches the hypercube and the whole space to generate a smaller elite hypercube until a termination criterion is met. Three popular benchmark problems in recent publications (such as IEEE Transactions on Evolutionary Computation and IEEE Transactions on Systems, Man, Cybernetics) and two popular benchmark problems at 2005 IEEE Congress on Evolutionary Computation are used for performance evaluation. HIA finds optimal solutions to the 5 benchmark problems. Under various conditions (population sizes, numbers of loops, and execution times), sufficient simulations indicate that HIA is more effective and more efficient than GA and the Bees Algorithm (BA). In the future, HIA will be improved, and a new HIA for discrete function optimization will be developed.