One-dimensional Cutting Stock Problem Based On Artificial Fish Swarm Algorithm

The one-dimensional cutting problem belongs to the combination and optimization problem and NP-hard. Looking for an optimum cutting solution can not only save raw materials and decrease produce cost, but also is an efficient method to improve the utilization of materials and to increase the benefit of enterprises. The one-dimensional cutting problems are usually solved using branch-and-bound, dynamic programming or integer linear programming. When dealing with the cutting stock problems with both mass material and parts, experts try to using the artificial intelligence algorithms like genetic algorithm and achieved better results. The artificial fish swarm (AFSA) is a new stochastic optimization algorithm based on group intelligent, essentially is a complex intelligent system. It has no special requirement for object functions, being insensitive to the initial values, tolerating wide range of values of parameters.The paper focuses on the improvement of AFSA algorithm and solving the one-dimensional cutting problem with the AFSA. The main achievements include:1) After analyzing the disadvantage of AFSA, an improved artificial fish swarm algorithm is presented. Firstly, it dynamically adjusts the vision of artificial fish. The way improves the global searching ability and increases the searching speed of the algorithm. Secondly, the preying behavior and the swarming behavior are improved in the algorithm. The simulation results show that this method is an efficient algorithm for solving optimization problems of function.2) In the paper, for solving the one-dimensional cutting problem with the AFSA, some parts of the AFSA are adjusted and redefined: the method of coding and initialization are introduced; the concepts of distance and center are redefined; the preying behavior, the swarming behavior and the following behavior are adjusted. Finally, the AFSA is applied to solve the one-dimensional cutting problem. The results demonstrate that the method is feasible and has the feature of high rate of convergence and high optimizing precision.