Distributed and Decentralized State Estimation Applied to Complex Epidemic Networks

Abstract

In epidemic models, the traditional approaches assume that each individual has an equal chance, per unit of time, to communicate with each other. In this regard, the use of complex networks can be considered a more realistic approach. Indeed, in epidemic complex networks the contact patterns are taken into account in the study of an epidemic spreading. However, due to the high dimensionality of the state and observation equations, the use of a classical centralized strategy for state estimation is a challenge. For this reason, as a preliminary study, we propose the use of distributed and decentralized information filter in order to overcome this issue. In a numerical example, for a susceptible-infected network, we show that the distributed and decentralized information filter is effective in state estimation as the centralized approach.