Abstract
This work presents a new global optimization algorithm based on the behavior of orb-web spiders. The spider method is a heuristic competitive iterative method of random search whose main idea is to model the behavior of Garden orb-web spiders. The paper describes a solution search strategy based on the features of building a web and competitive behavior; the step-by-step algorithm for solving the problem is presented. The concept of the method is simple; the algorithm does not contain complex calculations. The positions of spiders and flies (test points) are generated randomly, herewith, due to the correct selection of the obtained values, the search for an optimum lead to a rather accurate result. The article describes all the parameters used in the method and presents recommendations for changing settings. The selection of the optimal parameters for various classes of test functions is performed. Parameter settings were performed on three classes of test functions: unimodal, ravine, and multiextremal. As part of the work, a computational experiment was conducted to study the effectiveness of the developed method as well. We compared the proposed method with other proven metaheuristic optimization algorithms. The method showed good results both when working with simple (unimodal) functions, and when finding the optimum of functions with a more complex landscape (multiextremal functions). Based on the above advantages, the spider method can be adapted to solve applied problems with relative ease.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have