A System of Systems for the Optimal Allocation of Pollutant Monitoring Sensors

In this article, the problem of pollutant monitoring sensor allocation is formalized as a variant of the well-known set covering problem and solved by means of both a mixed-integer linear programming solver and genetic algorithms. The allocation of the sensors in the monitoring network is one of the most practical and challenging problems that regional authorities have to face, in particular due to the constraints that European Union Directives pose in terms of the number of sensors per area/subarea and in terms of representativeness capability (i.e., the ability that the data coming from the monitoring network have to represent the air quality levels in their surrounding areas). The methodology has been applied to the allocation of sensors for particulate matter (PM10) and nitrogen dioxides (NO$_2$) in Lombardy, Italy. The results are promising, showing that the two proposed approaches are able to strongly dominate the current solution.