Shared parking problem: A novel truthful double auction mechanism approach

Considering a parking platform with multiple slot demanders and multiple slot suppliers, this paper addresses two truthful double auction mechanisms for shared parking problem, where the demanders targeted by the proposed mechanisms are with flexible schedules and for non-work activities, and the suppliers are with the typical daily 'driving go out early and come back at dusk' pattern. To provide flexible match schemes and increase the trading scale, we allow all-or-nothing principle for demanders and one-to-many principle for suppliers. Based on the parking slot allocation rule and the transaction payment rule, we first relax the single output restriction (suppliers can only submit bids on a single unit of one commodity) and propose a demander competition padding method (DC-PM) auction mechanism, which is used to solve potential budget deficit. To avoid the likely distorted social welfare resulted by the DC-PM auction mechanism and to add budget surplus, we further modify the parking slot allocation rule and the transaction payment rule in the DC-PM auction mechanism, and propose a modified demander competition padding method (MDC-PM) auction mechanism. Three economical properties of both auction mechanisms, i.e. incentive compatibility, individual rationality and budget balance, are proved, Numerical experiments show that the proposed two auction mechanisms can realize asymptotic efficiency as both demanders and suppliers approach infinity. Moreover, the DC-PM auction mechanism is superior to the MDC-PM auction mechanism with respect to participants' utilities and can maintain the strict lexicographic allocation order, and the MDC-PM auction mechanism outperforms the DC-PM auction mechanism in terms of welfare efficiency and platform's payoff. Issues of managerial implications for shared parking problem are further discussed in this paper. (C) 2018 Elsevier Ltd. All rights reserved.