EvoBank: An evolutionary game solution for Bankruptcy problem

Hamidreza Mahini | Hamid Navidi Ghaziani | Fatemeh Babaei | Seyyedeh Mobarakeh Mousavirad

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=158729
Date :  2021/08/30
Publish in :    Swarm and Evolutionary Computation
DOI :  https://doi.org/10.1016/j.swevo.2021.100959
Link :  http://dx.doi.org/10.1016/j.swevo.2021.100959
Keywords :Game theory, Bankruptcy problem, Evolutionary game

Abstract :

---

There has long been a debate surrounding the bankruptcy problem and how to share bankrupt entity inventory among its claimants. In this paper, we propose a novel inventory sharing algorithm that is based on a population game and evolutionary dynamics. In contrast with most existing sharing rules, In our approach, all claimants are involved in deciding on participating in a game. We set up a population game in which its players corespond to claimants, and available sharing rules are their actions. Then we find the asymptotically stable state of the game or Nash Equilibrium based on pairwise proportional imitation (replicator dynamics). Finally, we propose the founded equilibrium profile in the mentioned game as a novel sharing rule. According to the rationality principle in game theory, there is no profitable individual deviation from any claimants’ equilibrium profile. So, the NE profile is a self-enforcement agreement between all claimants. The implementation of our proposed algorithm in MATLAB confirms its effectiveness completely