Human-Inspired Algorithms for continuous function optimization

The Human-Inspired Algorithm (HIA) is a new algorithm that uses a given population (a group of candidate solutions) to improve the search for optimal solutions to continuous functions in different optimization applications such as non-linear programming. HIA imitates the intelligent search strategies of mountain climbers who use modern techniques (such as binoculars and cell phones) to effectively find the highest mountain in a given region. Different from Genetic Algorithms (GAs) and Bees Algorithms (BAs), HIA divides a whole search space into multiple equal subspaces, evenly assigns the population in the subspaces, finds an elite subspace with the largest sum of function values, and uses more climbers (candidate solutions) to explore the elite subspace and fewer ones to explore the rest of the whole search space. BAs use random search in local neighborhood search, whereas HIA uses GAs in local neighborhood search to obtain better results. HIA locates a point with the largest function value among the elite sites and creates a hypercube with the point as its center. The assigned climbers in the hypercube and the elite subspace continue to search for the optimal solution iteratively. In each loop, the hypercube and the elite subspace become smaller to have a larger chance to pinpoint the optimal solution. Simulation results for three continuous functions with constraints and three continuous functions with box constraints can indicate that HIA is more efficient than GAs and BAs. Finally, conclusions and future works are given.