Multiobjective synthesis of robust vaccination policies

Authored by Cruz Andre R da, Rodrigo T N Cardoso, Ricardo H C Takahashi

Date Published: 2017

DOI: 10.1016/j.asoc.2016.11.010

Sponsors: Brazilian Ministry of Education (CAPES) Brazilian National Council for Scientific and Technological Development (CNPq)

Platforms: No platforms listed

Model Documentation: Other Narrative Pseudocode Mathematical description

Model Code URLs: Model code not found

Abstract

This paper deals with the optimal planning of vaccination campaigns, using an evolutionary multiobjective optimization algorithm and a stochastic simulation of the epidemics dynamics in order to determine robust vaccination policies. A biobjective model is formulated, considering the minimization of control costs and number of infected individuals. The decision variables include number of campaigns, percentage of vaccination and time interval between each campaign. A SIR (Susceptible-Infected-Recovered) model and an IBM (Individual-Based Model) are employed for representing the epidemics. A two-stage optimization process is proposed: a set of nondominated steady-state regimes is obtained and one of them is selected to be concatenated to the transient regime vaccination policies. An evolutionary multiobjective optimization algorithm is proposed, with a local search procedure based on quadratic approximation supported by a hash table information storage. The resulting nondominated solutions are simulated in the IBM, in order to detect and discard the non-robust solutions. Final results show that optimal robust vaccination campaigns with different trade-offs can be designed, allowing policymakers to choose the best strategy according to the monetary cost and the expected efficacy. (C) 2016 Elsevier B.V. All rights reserved.

Tags

Genetic Algorithms algorithms Optimization Strategies Diseases Sir epidemic model Multiobjective optimization Vaccination planning Robust synthesis Epidemics control