Spatial Assignment Optimization of Vaccine Units in the Covid-19 Pandemics

In this work we present the application of evolutionary algorithms to the problem of spatial assignment optimization of vaccine units. In the framework of urban planning of health facilities, the problem consists into optimizing the overall cost of building and running vaccine units with respect to costs and benefit for the public by deciding their size and location. The complex non linear objective function, depends on populations distributions, transportation infrastructure costs, travel times and distances and vaccination units capacities. The problem domain is described by a model based on a layered approach, where the layers embed knowledge of different types at a scalable resolution. Although many purposely designed algorithms for spatial locations assignment of health facilities, have been proposed in the literature, in a pandemics situation, for vaccination units, faster optimization tools are needed not necessarily designed for a specific problem model, which can quickly change dynamically. We have investigated and compared the application of several evolutionary optimization algorithms from PSO to Differential Evolution. Results show that evolutionary algorithms allow an high degree of flexibility in objective function without compromising in optimisation performance.