Envy-minimizing pareto efficient intersection control with brokered utility exchanges under user heterogeneity

- A signalized traffic intersection is controlled as a multi-agent exchange economy.
- The proposed system is Pareto optimal with regard to delay minimization.
- Novel envy-free based pricing guarantees maximal fairness and social acceptability.
- The system is financially self-sustainable and arbitrage is prevented.
- The method is computationally efficient and suitable for real-world implementation.

We propose PEXIC (Priced EXchanges in Intersection Control), a new concept and algorithm for traffic signal control that incorporates user heterogeneity on value of delay savings. The algorithm assigns phases with associated delays, taking into account the vehicle travelers' values for experienced delay. Applying principles of envy-freeness, we develop a pricing scheme that addresses fairness by minimizing user envy via compensatory monetary transfers among users. PEXIC is Pareto efficient and budget balanced, and thus financially self-sustainable without external subsidy. The optimization is solved sequentially on a rolling horizon basis: first the phasing, and next the pricing. PEXIC achieves significant cost reductions for a large range of volumes and users' value heterogeneity levels. Inclusion of user heterogeneity also proved to be fairer than standard delay minimization that disregards individual vehicles' values for delay savings. Furthermore, we show that arbitrage is not possible, thus there are no incentives to drive just to collect those payments. The method used has polynomial complexity and it is suitable for real-world implementation.