Algorithm for solving the problem of irregular pattern cutting using of artificial neural networks

Abstract

The article is devoted to the problem of placing non-standard geometric shapes on a rectangular sheet with the condition of maximum filling density. This task belongs to the class of irregular cutting-packaging and meant to minimize waste after cutting sheet material. The article describes the formulation of the problem of cutting sheet material. The parameters influencing the solution of the cutting problem are identified. The restrictions imposed on the implementation apparatus are determined. A brief analysis of existing methods suitable for solving the problem of location and spacing of blanks sheet material is carried out. The features of the operation of neural networks that affect the solution of the problem are specified. The choice of a neural network model with a description of the necessary mathematical apparatus for solving the cutting problem are made. A new combined algorithm for solving the problem of cutting sheet material based on the technology of artificial neural networks is proposed. A procedure for training an agent with a formula for his rewarding is introduced. The results of experimental studies of a software product based on the proposed algorithm are presented.