Enemy Surrounding Inspired Optimisation Algorithm: Introduction and Tests

Nature-inspired metaheuristic algorithms are becoming more popular for solving optimization problems because they offer many advantages over traditional numerical optimization techniques. The capacity of nature-inspired algorithms to discover answers in a short amount of time is their most significant benefit. Their speed is one of their most advantageous characteristics, yet it is also one of their biggest liabilities. As a result of the increased speed of operation, there is a possibility of subsidence in the local optimum, the discovery of single solutions for optimization problems containing more than one global minimum, difficulty in replicating the achieved results, and complicated configuration due to the large number of control parameters.

This paper proposes an optimisation algorithm called Enemy Surrounding Inspired Optimization (ESIO) Algorithm  and applies it to challenging tests in optimisation. Prepared algorithm, in contrast to the rest of its field, was not inspired by nature, but rather by a mix of well-known numerical technique and method based on the concept of an insect swarm. The algorithm was inspired by the ancient battlefield, during which warriors search opposing forces at various speeds based on the weapon they are wielding and then restrict the scope of the battlefield by firing at regular intervals at equal distances. The assumptions on which the proposed method is based are similar. One of the major assumptions of the study was to show the potential benefit from a combination of optimization techniques and to compare the mixed approach with methods based on the concept of an ad hoc swarm in terms of operation speed and efficacy in reaching the global optimum.

The method was evaluated in comparison to two prominent metaheuristic algorithms in terms of the accuracy of the answers discovered, the speed of the algorithm, and the simplicity of use of the algorithm. In accordance with the findings of the tests, it has a comparable performance to other algorithms and produces acceptable results when put into practice.