Site selection by Monte Carlo method and integration with brute-force search and genetic algorithm by using image processing approach (Case study: Fuel station in Tabriz city)

The purpose of this research is to find the optimal location for establishing a new service unit, like fuel station, within the urban area so that the average distance traveled by each user to the nearest unit reaches the lowest value. For this purpose, Monte Carlo simulation method was applied along with two optimization approaches including brute-search and genetic algorithm. In addition, image processing tools were used for distinguishing the urban regions and identifying the border of each zone. As a case study, the construction of a new fuel station unit in Tabriz city is carried out in this article. To achieve this goal, 40,000 users were randomly selected according to the population density of each zone within the city and the average distance of each of them was calculated from the nearest station. Afterwards, using the two mentioned algorithms, the new station was added in such a way that the minimum mean traveling distance was achieved. Considering the same number of random users for both methods, the genetic algorithm with initial population, generation and mutation rate of 60, 30 and 0.2, respectively, was resulted in better performance in terms of time and the mean distance. The mean distance between drivers and stations before adding a new unit was 2105 meters. However, by adding a new fuel station using the brute-search method and the genetic algorithm, this distance is reduced to 1908 and 1901 meters, respectively.