The Magic of Nash Social Welfare in Optimization: Do Not Sum, Just Multiply!

In this paper, we explain some key challenges when dealing with a single/multi-objective optimization problem in practice. To overcome these challenges, we present a mathematical program that optimizes a Nash Social Welfare function. We refer to this mathematical program as the Nash Social Welfare Program (NSWP). An interesting property of the NSWP is that it can be constructed for any single/multi-objective optimization problem. We show that solving the NSWP could result in more desirable solutions in practice than its single/multi-objective counterpart. We also discuss several promising approaches that could be employed to solve the NSWP in practice.

Article

Download

View PDF