| Seagull optimization algorithm establishes an algorithm model by simulating seagull migration and attack behavior,which has the characteristics of simple structure and easy understanding.However,the algorithm has some shortcomings,such as low optimization accuracy,slow convergence speed and the original algorithm can not deal with multi-objective optimization problem.In this paper,the shortcomings of seagull optimization algorithm are analyzed and improved,and the proposed algorithm is applied to UAV path planning.The main research contents are as follows:(1)Seagull optimization algorithm(SOA)is slow in convergence and easy to fall into local optimum.To solve these problems,a quantum-inspired seagull optimization algorithm(QSOA)is proposed.The initial position of the seagull population is encoded by qubits,the wave function composed of the message-encoded qubits is used to calculate the optimal position of the seagull population,and a variable angular-distance rotation(VAR)gate is used to change the probability amplitude of the qubit,thereby updating population.Experiments on a set of benchmark functions show that QSOA has better optimization ability and convergence speed than other eight heuristic algorithms.Finally,the algorithm is applied to the three-dimensional path planning problem of UAV.The results show that QSOA can find the shortest path and converge faster than other similar algorithms.(2)In practical application,there are not only single-objective optimization problems,but also multi-objective optimization problems,but the original SOA can not deal with multiobjective optimization problems.Inspired by this,a quantum-inspired multi-objective seagull optimization algorithm based on decomposition(MOQSOA/D)is proposed.The algorithm transforms the multi-objective problem into multiple scalar optimization subproblems,and establishes dynamic archive and leader archive at the same time.The Pareto solution of each sub-problem is stored in a dynamic archive,and the non-dominated Pareto solution is stored in the leader archive.When dealing with each subproblem,the seagull population is coded by qubits to control the current optimization direction.Penalty-based boundary intersection approach is introduced to determine whether the generated Pareto solution is retained.The proposed algorithm is tested with six different algorithms on 57 test functions,and the experiment shows that the algorithm achieves 30 better results.In addition,the algorithm is applied to the multi-objective path planning model with risk matrix,and the results show that MOQSOA/D obtains more dominant results in Pareto solution than the other algorithm. |
