Optimization of the Containment Levels for the Reopening of Mexico City due to COVID-19
Keywords:COVID-19, rein, deep q-learning, Genetic Algorithms, model simulation
One of the main problems that governments face in a COVID-19 pandemic is ensuring public health of a country while reducing the economic effects. There is an implicit trade-off between economic wealth and COVID-19 cases and deaths. As less restrictions are imposed on citizens to minimize economic losses, more difficult it is to stop the spread of the virus. Conversely, many restrictions to reduce the pandemic carry out a huge damage to the economy. Some researchers have proposed models to plan governments actions in different countries, but very few consider a necessary balance between public health and economy. Given the high population density of Mexico City, it is particularly urgent to consider a way to support decision-making considering the balance between the two aspects mentioned. In this paper, we propose a methodology to help the government plan the number of days for each contingency phase applied in Mexico City. This methodology includes a health model, an economic impact model when restrictive actions are taken, and optimization methods to ponder health and economic aspects in a society. We propose two scenarios in which the phases are predicted for the next 120 days. The first one seeks that the number of beds will never be exceeded considering less relevant the economic impact. The second experiment reduces the number of days in which beds are exceeded as long as the economic losses are not higher than 20%. Our methodology uses a SEIR model to simulate the evolution of the pandemic and two optimization algorithms: genetic algorithms and reinforcement learning.