A quality assessment for solutions achieved through algorithmic variations of the Multi-Start method

Abstract

For this research, an experiment with two algorithmic variations of the Multi-Start heuristic method for local search was held in order to generate and assess the quality of solutions coming from different manufacturing scenarios. These variations were applied to 29 oil refinery models to infer about the effectiveness and convenience of each model variation. A commercially available software, frequently used in blending problems at the industrial level, was used.