Efficient Utilization of Metaheuristic Techniques for Profit Maximization in Combinatorial Production Planning: A Comparative Study

Efficient production planning requires well-structured models compatible with metaheuristic and other optimization techniques. This study compares three established metaheuristic methods - sanitized-teaching-learning-based optimization (s-TLBO), Particle Swarm Optimization (PSO), and Differential Evolution (DE) with a termination criterion of equal number of functional evaluations - for profit-maximizing production planning. The focus is on a single-level problem involving complex constraints, non-linear costs, budget and raw material limitations. The paper presents two solution strategies: a penalty-based and a correction-based approach. Analyzing 300 unique instances of the problem reveals that s-TLBO implemented with correction-based approach consistently outperforms the other methods in devising optimal production plans for two Saudi Arabian petrochemical industries with 24 products, 54 processes, and three raw materials for plant 1 and two raw materials for plant 2.