An approach that integrates artificial neural networks and optimization methods for automating regularly-shaped pattern generation processes is proposed. In the proposed approach the artificial neural network model is used for generating rectangular pattern configurations with an acceptable scrap. Rectangular patterns of different sizes are used as the input of the neural network to generate location and rotation of each pattern when they are combined. The pattern configurations generated through the neural network are represented as decision variables of a mathematical programming model for determining an efficient nesting of different sizes of rectangular patterns for meeting the demand of a given planning horizon. The initial results obtained based on rectangular patterns are reported

Meeting Name

International Joint Conference on Neural Networks, 1992


Engineering Management and Systems Engineering

Keywords and Phrases

CAD; Computational Geometry; Nesting; Neural Nets; Neural Network; Optimisation; Optimization; Planning Horizon; Rectangular Pattern Configurations; Rectangular Patterns; Regularly-Shaped Pattern Generation

Document Type

Article - Conference proceedings

Document Version

Final Version

File Type





© 1992 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.