الگوریتم بهینه سازی متا هیوریستیک جدید با الهام از شکار گروهی از حیوانات: جستجوی شکار

A novel optimization algorithm is presented, inspired by group hunting of animals such as lions, wolves, and dolphins. Although these hunters have differences in the way of hunting, they are common in that all of them look for a prey in a group. The hunters encircle the prey and gradually tighten the ring of siege until they catch the prey. In addition, each member of the group corrects its position based on its own position and the position of other members. If the prey escapes from the ring, hunters reorganize the group to siege the prey again. Several benchmark numerical optimization problems, constrained and unconstrained, are presented here to demonstrate the effectiveness and robustness of the proposed Hunting Search (HuS) algorithm. The results indicate that the proposed method is a powerful search and optimization technique. It yields better solutions compared to those obtained by some current algorithms when applied to continuous problems.

Classical methods often face great difficulties in solving optimization problems that abound in the real world. In order to overcome the shortcomings of traditional mathematical techniques, nature-inspired soft computing algorithms have been introduced. Several evolutionary or meta-heuristic algorithms have since been developed which combine rules and randomness mimicking natural phenomena. These phenomena include biological evolutionary processes (e.g., the evolutionary algorithm proposed by Fogel et al. [1], De Jong [2], and Koza [3] and the genetic algorithm (GA) proposed by Holland [4] and Goldberg [5]), animal behavior (e.g., the tabu search proposed by Glover [6]), the physical annealing process (e.g., simulated annealing proposed by Kirkpatrick et al. [7]) and the musical process of searching for a perfect state of harmony (proposed by Geem et al. [8], Lee and Geem [9] and Geem [10] and proceeded with other researchers [11] and [12]). Many researchers have recently studied these meta-heuristic algorithms, especially GA-based methods, to solve various optimization problems. However, new heuristic algorithms are needed to solve difficult and complicated real-world problems. The method introduced in this paper is a meta-heuristic algorithm which simulates the behavior of animals hunting in a group (lions, wolves, etc.). Group hunters have certain strategies to encircle the prey and catch it. Wolves, for instance, rely on this kind of hunt very much, so they can hunt animals bigger or faster than themselves. They choose one prey and the group gradually moves toward it. They do not stand in the wind such that the prey senses their smell. We employ this idea in constrained problems to avoid forbidden areas. In our algorithm, each of the hunters indicates one solution for a particular problem. Like real animals which hunt in a group, artificial hunters cooperate to find and catch the prey; i.e., the optimum point in our work.

A new meta-heuristic algorithm has been developed. The Hunting Search (HuS) meta-heuristic optimization algorithm was conceptualized using the strategy of group hunters in catching their prey. Compared to gradient-based mathematical optimization algorithms, the HuS algorithm imposes fewer mathematical requirements and does not require initial value settings of the decision variables. In addition, the HuS algorithm uses stochastic searches; therefore, derivative information is unnecessary. Selected benchmark optimization problems were solved to demonstrate the effectiveness and robustness of the new algorithm compared to other optimization methods. The test cases showed that the HuS algorithm is a global search algorithm that can be easily applied to various optimization problems. The results obtained using the HuS algorithm would yield better solutions than those obtained using other algorithms. In addition, in constrained optimization problems, hunters have power of orientation in the design space (movement toward leader); therefore, they can escape from forbidden areas and find feasible areas quickly, as real hunters do. Further work is still needed to solve more complex and real optimization problems such as engineering problems. The algorithm can also be generalized for solving discrete and combinatorial optimization problems such as the traveling salesman problem and timetabling.