A Mixed Linear Integer Programming Formulation And A Simulated Annealing Algorithm For The Mammography Unit Location Problem

Breast cancer is the most commonly occurring one in the female population. Early diagnosis of this disease, through mammography screening, can increase the chances of cure to 95%. Studies show that Brazil has a relatively satisfactory number of mammography units, but this equipment is poorly geographically distributed. This paper focuses on the Mammography Unit Location Problem (MULP), which aims an efficient distribution of mammography units, in order to increase the covered demand. Focusing on the State of Minas Gerais, Brazil, an analysis is made considering that, in the real world, there is a difficulty in relocating equipment already installed. Therefore, it would be interesting to optimize the location of new equipment purchases. Since MULP is NP-hard, an algorithm based on the Simulated Annealing meta-heuristic is also developed to handle large instances of the problem.