Dynamic Stochastic Matching Under Limited Time

In “Dynamic Stochastic Matching Under Limited Time,” Aouad and Sar?taç analyze the design of matching policies in dynamic markets such as carpooling platforms kidney exchange schemes. A crucial distinction with previous literature is that agents’ arrivals departures are fully dynamic. The demand supply side constantly replenished; each market participant remains available for potential matches during a limited period time. Specifically, authors formulate general model over edge-weighted graphs, where agents' abandonments stochastic heterogeneous. platform controls how long agent waits whom s/he matched with. These decisions subject to fundamental tradeoff between increasing thickness mitigating risk from certain participants. authors’ main contribution devise simple algorithms strong performance guarantees broad class networks. contrast, they show widely used batching have an arbitrary bad on graph-theoretic structures. Their analysis involves novel techniques including linear programming benchmarks, value function approximations, proxies continuous-time Markov chains, which may be broader interest. Extensive simulations real-world taxi data demonstrate newly developed can significantly improve cost efficiency against algorithms.