Differential Evolution Based Simulated Annealing Method for Vaccination Optimization Problem

Infectious diseases pose a severe threat to human health, especially the outbreak of COVID-19. After the infectious disease enters the stage of large-scale epidemics, vaccination is an effective way to control infectious diseases. However, when formulating a vaccination strategy, some restrictions still exist, such as insufficient vaccines or insufficient government funding to afford everyone's vaccination. Therefore, in this paper, we propose a vaccination optimization problem with the lowest total cost based on the susceptible-infected-recovered (SIR) model, which is called the Lowest Cost Of Vaccination Strategy (LCOVS) problem. We first establish a mathematical model of the LCOVS problem. Then we propose a practical Differential Evolution based Simulated Annealing (DESA) method to solve the mathematical optimization problem. We use the simulated annealing algorithm (SA) as a local optimizer for the results obtained by the differential evolution algorithm (DE) and optimized the mutation and crossover steps of DE. Finally, the experimental results on the six data sets demonstrate that our proposed DESA can achieve a more low-cost vaccination strategy than the baseline algorithms.