Multi-Objective Optimization Application Research Based On Multi-Modal Lion Swarm Algorithm

Heuristic algorithms are an important branch of artificial intelligence research,a class of algorithms developed based on natural phenomena or mechanisms inspired by them,widely used in black-box optimization,combinatorial optimization and multi-objective optimization,and have become the mainstream method for solving complex optimization problems.In this paper,we focus on one branch of heuristic algorithms:the lion swarm optimization algorithm.The multimodal and diverse behavioral mechanisms of the lion swarm optimization algorithm enable it to search for optimal solutions more comprehensively,reduce the probability of falling into local optimal points,and better adapt to different situations.In this paper,based on the lion swarm optimization algorithm,different heuristic algorithms are combined to form a multi-modal lion swarm algorithm,and the multi-objective optimization problem is studied as follows:(1)To address the problem that the single objective changes with time interval,which affects the timely finding of the global optimal solution in time variation.A multimodal small habitat lion population optimization algorithm based on sensitive particles to identify dynamic environments is proposed.The method generates a dynamic environment that changes with time through the dynamic function generator DF1 to realize the multimodal function.And sensitive particle recognition is introduced to find the optimal function for the current environment.The method improves the recognition tracking ability and adaptation ability.Meanwhile,the comparison experimental results show that the method has higher recognition ability and tracking ability for the changing state of the environment compared with the particle swarm optimization algorithm.(2)To address the problem that the multi-objective lion swarm algorithm is prone to local over-concentration and poor search performance when the dimensionality is high,and it is difficult to maintain the population convergence and diversity,the quantum computing is referred to,and the real number encoding mechanism as well as the quantum gate update mechanism are introduced in the algorithm to adjust the population with double chains,and the quantum gate update is applied to increase the diversity.The performance improvement of the quantum multi-target lion population algorithm is verified through simulation experiments on the benchmark test problem.The application area of the quantum multi-objective lion swarm algorithm is further extended by applying quantum multi-objective lion swarm to solve a multi-objective scheduling model for the economic and environmental benefits of grid-connected wind power.The experimental results show that the quantum lion swarm multi-objective algorithm is feasible for solving the scheduling multi-objective problem.(3)A multi-objective lion swarm algorithm based on teaching and learning is proposed by combining teaching and learning optimization algorithm and lion swarm optimization algorithm.By using the mechanism of teacher-led classmates’ performance improvement in the teaching-learning theory,the drawback of the original lion swarm algorithm that cannot approximate the real frontier surface of multi-objectives is effectively overcome,and the diversity and accuracy of the algorithm are improved.It is also applied to the logistics distribution center site selection problem for experimental validation.The experiments show that the improved algorithm outperforms the traditional lion swarm algorithm in multi-objective optimization in high-dimensional conditions and shows better performance in real-world application problems.