Multi-objective optimization of Gas Station performance using response surface methodology

Purpose: In this paper, the performance of a specific gas station (Parvin) has been studied and investigated. The purpose of this research is to design a second-order regression model based on simulated data to optimize the queuing system in line with the fuel sales and costs. Thus, the influential variables including the number of pumps and the number of pump operators need to be optimally determined. Design/methodology/approach: The simulation was combined with design of experiments (DoE) techniques to achieve a predictable model for optimizing Gas Station performance considering both the sales rate and the queue length. First, the Gas Station was simulated with Arena software, and then by using DoE and response surface methodology (RSM), the gas station performance was optimized in terms of three objectives including costs. A face-centered central composite design (CCD) has been implemented to reach the optimal number of pumps and pump workers. Findings: The results of the optimization model derived from the CCD indicate that the performance of the Gas Station system has been improved considerably. Moreover, after the detailed study of optimization and RSM outputs, it seems that the variations of both the pumps and the number of pump operators have significant impacts on the performance of the Gas Station including costs, sales rate and queue length. Originality/value: In general, it has been proved that simulation-based RSM can be considered as a powerful and effective technique in both single and multi-objective experimental optimization. The present study has been able to help managers to make decisions and conduct the Gas Station in critical conditions in different hours of a day.