An Intelligent Approach to Stock Cutting Optimization

The stock cutting problem has gained a lot of attention as a means to increase efficiency in many industrial sectors. In this paper, we present a multi-level, hierarchical, architecture exploiting different artificial intelligence methodologies for stock cutting optimization. An expert system is used at the highest level for decision functions; the second layer utilizes genetic algorithms for pattern construction. A third layer using linear programming for pattern selection follows. At the lowest layer, a tool set for computer implementation uses heuristic and maintains constraints. A prototype software package based on this architecture has been implemented and tested on a windows based workstation using visual C++. Sample output results demonstrating different features are presented.