Location analysis for a grocery store based on a multi-objective optimization approach

In this paper, we propose a method allowing to find the optimum location for a grocery store in an urban area. In our algorithm, we use a multi-objective optimization approach where for a given geographic area, we extract a set of solutions based on two criteria: The first one attempts to minimize the distance from places like restaurants, bus stations, etc. as these places denote a pedestrian traffic. The second one tries to maximize the distance from other existing grocery stores, in order to find a location with less competition. The multi-objective genetic algorithm (MOGA) utilized proposes a set of solutions that cannot dominate each other. Therefore, for the geographic area analyzed by MOGA, after detecting the surfaces corresponding to buildings based on color map information, we calculate the average weighted mean of the building surfaces. Hence, we are selecting among the solutions proposed by MOGA the closest one to the calculated weighted mean, in an effort to be located near a dense population. After testing the system with different scenarios, we show that this application is able to propose adequate locations in respect to the predefined criteria.