Genetic Algorithm Approach for Cutting Stock Problems in Construction Industries

Abstract

Cutting Stock Problem (CSP) is a classic problem involving cutting long stocks into smaller products with certain quantities. The optimization is to find cutting patterns with minimum waste. In construction industries, CSP applies to steel bar cutting. The steel bar is an important element in making reinforced concrete. The length of the steel bars from the steel manufacturers is fixed, while the requirements for the constructions are varying. The problem is to find optimized way to cut long, fixed-length steel bars into smaller, varying length bars required in the constructions. The requirements are different from building to building, both in the lengths and quantities. Many studies have been extensively done on the subject, from Brute Force, Greedy Search to Linear Programming. In this paper the study focuses on Genetic Algorithm approach. The results look promising for Fitness Function 1 where the focus is to minimize waste. Waste ranges from 2.03% to 4.31%. Fitness function 2 and 3 do not emphasize merely on minimizing the waste, but also on contiguity. Therefore the residues are more, ranges from 2.21% to 4.91% for Fitness Function 2 and from 2.03% to 30.7% for Fitness Function 3