Determining best grocery store location based on a multi-objective optimization approach

Multi-objective optimization is an approach that can be utilized to optimize more than one objective function simultaneously. NSGA-II is a non-dominated sorting genetic algorithm that has been widely used for solving different multi-objective optimization problems. Location decision is a critical problem for the facilities such as markets and stores. The goal of this study is to decide where a new facility should be located. In the literature, there are studies that utilize multi-objective optimization techniques for the location decision problem. However, such applications try to determine the location for big facilities like hospitals, big grocery stores and so on. In this study, a novel approach based on multi-objective genetic algorithms (MOGAs) have been proposed for the location decision problem of moderate-sized grocery stores. The genetic search is also supported with image processing techniques in a hybrid framework and the NSGA-II is utilized as the multi-objective genetic algorithm. The proposed method optimizes two objectives aiming to minimize the distance to restaurants and metro stations in the area and maximize the distance to other markets in the area. The mean value of the surface areas, an important factor for the location decision is found with Google Maps using image processing techniques that have the goal to detect the surface of individual buildings. The mean value of the surface areas are utilized to determine the final result among the set of solutions produced by NSGA-II. Various tests are carried out and it has been observed that the system can propose appropriate locations based on the objective functions utilized.