Genetic Algorithm For Solving One-dimensional Cutting Stock Problems

In this thesis, genetic algorithms are studied from coding. On the basis of the study genetic algorithms are applied to solving one-dimensional cutting stock problems. Then several genetic algorithms were designed to solve different cutting stock problems. Some good numerical results are reported.Firstly, in this thesis, we discuss the development of one-dimensional cutting stock problems and other well-known algorithms about them, summarize the basic principle of genetic algorithms, and analyze the effect of coding, fitness function, crossover operators and mutation operators in the genetic process of genetic algorithms.Secondly, cutting stock problems were studied systematically, which were solved by genetic algorithms. Numerical sign coding is improved to solve one-dimensional cutting stock problems. According to this coding we put forward genetic algorithm with Elitist Model to make new algorithm more effective. At the same time, crossover and mutation are developed to construct genetic algorithm of one-dimensional cutting stock problem.Finally, two new genetic algorithms are applied to solve the cutting stock problem. The results demonstrate that new genetic algorithms have predominance in the general cutting stock problems.