Treffer: A novel strategy for the combinatorial production planning problem using integer variables and performance evaluation of recent optimization algorithms.

Abstract In this article we propose a multi-unit production planning based optimization strategy that employs a set of integer and continuous variables to overcome many of the drawbacks of the formulation/strategies in literature [1,2] and helps in determining efficient production plans. Additionally, we have utilized an efficient strategy to handle the domain-hole constraints that does not rely on imposing penalties for violation. The benefits of the proposed strategy are demonstrated on eight cases, which have been previously discussed in literature to potentially guide the petrochemical industries, and provide an improvement of up to 34.66% in the profit. In addition to proposing an efficient strategy, this work also discusses the computational performance of five-optimization algorithms viz., the recently proposed sanitized-teaching-learning-based optimization algorithm, multi-population based ensemble differential evolution, enhanced binary particle swarm optimization algorithm, artificial bee colony algorithm and the popular real coded genetic algorithm on this combinatorial optimization problem. [ABSTRACT FROM AUTHOR]