Online stochastic weighted matching algorithm for real‐time shared parking

Abstract Parking management plays an important role in the efficiency of transportation systems. We study the problem of matching shared parking slots with demanders online on a parking platform with the aim of managing parking spaces intelligently and efficiently. In our setting, the demanders submitting parking requests sequentially are assumed to be drawn from a distribution over demander types. No further assumptions about the demander arrival stream are needed. Matching decisions are triggered by arrival events instead of time points. This benefits the demanders by immediate responses to parking requests. An online matching algorithm is designed to make real‐time decisions and maximize social welfare. We demonstrated that the algorithm has a large competitive ratio, which implies a good performance of the algorithm. Numerical experiments are conducted to further examine the performance of the algorithm. The experimental results show that our proposed algorithm can achieve such a large competitive ratio in a real‐time matching way as the batch matching algorithm does in a serious delay matching way.