Incentives in Public Goods Games on Networks

Public goods games (PGGs) on networks are multiagent games where individuals can invest in a public good, and agents benefit from investments in their neighborhoods. Solutions of PGGs are stable states (e.g., pure strategy Nash equilibria) and it is established that they do not always exist. The present study proposes an iterative multi-agent search algorithm that is guaranteed to converge on general binary PGGs on networks. The algorithm uses payments among neighboring agents during search and enforces a stable state on strategic agents via a VCG mechanism. The algorithm is guaranteed to converge to outcomes that are at least as efficient as the initial state. Outcomes are stabilized by the VCG mechanism upon termination of the algorithm and this stabilization by the mechanism incurs a cost on society. The experimental evaluation shows that this cost is only a small fraction of the increase in social welfare. The experimental evaluation compares the performance of the proposed algorithm to former solving methods on instances of binary PGGs where former methods converge (i.e., potential games). The proposed algorithm finds solutions of higher efficiency and fairness than the solutions obtained by former methods.