基于人工鱼群算法的多目标背包问题研究-数学专业论文.docxVIP

广东工业大学硕士学位论文ABSTRACT 广东工业大学硕士学位论文 ABSTRACT Multi-objective optimization problem and knapsack problem have been difficult and hot issues in the field of science and engineering research．Compared with the single objective knapsack problems，multi—objective knapsack problem generally include two more optimization objectives，SO the problem will be more complicated．Classical optimization algorithm such as dynamic programming is difficult to feasible computational cost and time search with high quality solutions，SO we need to study a more efficient algorithm structure to quickly find the Pareto optimal solution． This paper sums up the common methods of two swarm intelligence algorithms for solving multi‘objective knapsack problem：Genetic algorithm and particle swarm optimization．Genetic algorithm calculation is simple，easy to achieve programming，but it is prone to premature phenomena and near optimal solutions in the vicinity of the optimal solution；Particle swarm algorithm is fast running，but it is of low accuracy．This paper introduces artificial fish swarm algorithm in detail，generalizes several commonly used distance and artificial fish swarm algorithm commonly used coding method，and focuses artificial fish swarm algorithm for solving multi-objective knapsack problem．At last， according to the code for artificial fish，artificial fish moving strategy,we design improved artificial fish swarm algorithm based on global artificial fish swarm algorithm． To solve the muki-objective knapsack problem，artificial fish swarm algorithm has blind search，high complexity,low accuracy and slow convergence rate．Knapsack problem generally binary coding to solve the problem，but the of binary encoding is required for frequent encoding and decoding will greatly increase the amount of computation algorithm．In the artificial fish swarm algorithm，the actual of distance between the two fishes is the Euclidean distance with blindness and randomness．In view of these problems， the main work of th