Expected Bipartite Matching Distance in A $D$-dimensional $L^p$ Space: Approximate Closed-form Formulas and Applications to Mobility Services (2406.12174v1)

Abstract: Although many well-known algorithms can solve the bipartite matching problem instance efficiently, it remains an open question how one could estimate the expected optimal matching distance for arbitrary numbers of randomly distributed vertices in a $D$-dimensional $L^p$ space (referred to as a random bipartite matching problem, or RBMP). This paper proposes an analytical model with closed-form formulas (without statistical curve-fitting) that estimate both the probability distribution and expectation of the optimal matching distance of RBMP. Simpler asymptotic approximations of the formulas are also developed for some special cases. A series of Monte-Carlo simulation experiments are conducted to verify the accuracy of the proposed formulas under varying conditions. These proposed distance estimates could be key for strategic performance evaluation and resource planning in a wide variety of application contexts. To illustrate their usefulness, we focus on mobility service systems where matches must be made between customers and service vehicles that are randomly distributed over time and space. We show how the proposed distance formulas provide a theoretical foundation for the empirically assumed Cobb-Douglas matching function for taxi systems, and reveal conditions under which the matching function can be suitable. Our formulas can also be easily incorporated into optimization models to select taxi operation strategies (e.g., whether newly arriving customers shall be instantly matched or pooled into a batch for matching). Agent-based simulations are conducted to verify the predicted performance of the demand pooling strategy for two types of e-hailing taxi systems. The results not only demonstrate the accuracy of the proposed model estimates under various service conditions, but also offer valuable managerial insights for service operators to optimize their strategies.