POEM Project and Solutions for One-dimensional Multi-criteria Cutting Stock Problems

Abstract

In this paper, we present an induatrial project and consider a real life unidimensional cutting stock problem. It consists of not only minimizing trim loss as in traditional cutting stock problems but also minimizing cutting time as well. A variety of constraints should be taken into account. These constraints often arise in industry cutting iron and steel materials. It is well known that cutting stock problems are NP-hard. It is prohibitive to obtain optimal solutions. We develop approximation algorithms for different purposes: quick reponse algorithms for individual customer planning and algorithms to provide production plan for the next day. These latter algorithms have less computation time restrictions. Our algorithms improve 8% the current performance of the company we work for.