Optimal control of information in social network using pontryagin's maximum/minimum principle / Visnu Varatan Arumugham

Visnu Varatan, Arumugham (2018) Optimal control of information in social network using pontryagin's maximum/minimum principle / Visnu Varatan Arumugham. Masters thesis, University of Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (557Kb)
Preview	PDF (Thesis M.A) Download (1364Kb) | Preview

Abstract

This study is to formulate an optimal control problem to maximize the spread of information’s on the constant budget. Advertisement in a social media is the control signals which attempt to convert ignorants and stiflers into spreaders in this study. We show the existence of a solution to the optimal control problem when the campaigning incurs non-linear costs under the isoperimetric budget constraint. The solution employs Pontryagin’s Minimum Principle and a modified version of forward backward sweep technique for numerical computation to accommodate the isoperimetric budget constraint. The techniques developed in this paper are general and can be applied to similar optimal control problems in other areas. We have allowed the spreading rate of the information epidemic to vary over the campaign duration to model practical situations when the interest level of the population in the subject of the campaign changes with time. The shape of the optimal control signal is studied for different model parameters and spreading rate profiles. We have also studied the variation of the optimal campaigning costs with respect to various model parameters. Results indicate that, for some model parameters, significant improvements can be achieved by the optimal strategy compared to the static control strategy. The static strategy respects the same budget constraint as the optimal strategy and has a constant value throughout the campaign horizon. This work finds application in election and social awareness campaigns, product advertising, movie promotion and crowdfunding campaigns.

Actions (For repository staff only : Login required)