An evolutionary approach to two-dimensional guillotine cutting problem

An evolutionary approach, based on genetic algorithms, is presented to solve the general guillotine cutting problem. The 2D cutting problem is the problem of cutting a single plane rectangle into a number of rectangular pieces of given sizes and values so as to maximize the total value of the pieces cut. Most of the previous approaches to tackle this problem were based on dynamic programming. We propose a binary tree representation to encode the problem space, together with some genetic operators to manipulate such structures. A number of test problems, including a real-world problem, have been used to illustrate the computational performance of our algorithm. Computational results are presented which clearly demonstrate the effectiveness of our approach