An integrated genetic algorithm approach to 1D-cutting stock problem

Manufacturing industries face trim minimisation problem, which if not effectively dealt results in loss of revenue. In this paper, we propose a new genetic-based approach to solve one dimensional cutting stock problem. The approach involves effective column generation techniques to stabilise and accelerate the solution process. This acceleration is achieved by imposing penalty function on the fitness value for evolution of better population. The GA capability is enhanced by using dynamic behaviour in crossover and mutation operators. The dynamism helps to improve the solution convergence rate to a great extend and controls the random behaviour to acceptable levels. Our approach reduces the rate by 60%. The computation comparison with the existing similar LP-based hybrid approach and other existing and recent meta heuristic approaches from literature proves the feasibility and validity of the algorithm. The proposed approach proves its efficiency and applicability on benchmark as well as industrial problems.