Solving the patient zero inverse problem by using generalized simulated annealing

Abstract

Identifying patient zero – the initially infected source of a given outbreak – is an important step in epidemiological investigations of both existing and emerging infectious diseases. Here, the use of the Generalized Simulated Annealing algorithm (GSA) to solve the inverse problem of finding the source of an outbreak is studied. The classical disease natural histories susceptible–infected (SI), susceptible–infected–susceptible (SIS), susceptible–infected–recovered (SIR) and susceptible–infected–recovered–susceptible (SIRS) in a regular lattice are addressed. Both the position of patient zero and its time of infection are considered unknown. The algorithm performance with respect to the generalization parameter q̃v and the fraction ρ of infected nodes for whom infection was ascertained is assessed. Numerical experiments show the algorithm is able to retrieve the epidemic source with good accuracy, even when ρ is small, but present no evidence to support that GSA performs better than its classical version. Our results suggest that simulated annealing could be a helpful tool for identifying patient zero in an outbreak where not all cases can be ascertained.