A two-stage budget-feasible mechanism for mobile crowdsensing based on maximum user revenue routing

Through user participation, mobile crowdsensing (MCS) services overcome the problem of the excessive costs of relying solely on the active deployment of sensors and of achieving large-scale and low-cost applications of the Internet of Things, which is a research hotspot. However, current research on MCS issues adopts the perspective of service providers and does not consider user strategies, so the corresponding models cannot accurately reflect the complete status of the system. Therefore, this paper decomposes the MCS problem into a two-stage game process. By doing so, the strategies of both users and service providers can be considered, thus maximizing the interest for both parties. In the first stage, users determine the optimal route based on information released by the service provider. In the second stage, the service provider determines the winning users and the corresponding payment plan based on the route and bid information submitted by all users. Specifically, we express the user’s optimal route decision-making problem as a traveling salesman problem with time windows and node number constraints. Accordingly, we design the F-MAX-RR algorithm based on an evolutionary algorithm. We show that this algorithm can achieve an approximation ratio of $(1-1/e)$, with the expected number of iterations being $8e{L}^2(L+1)M$. In the second stage, to maximize the total utility of the system, we transform the problem into an integer programming model with a budget constraint, which satisfies submodular characteristics. We design the S-MAX-TUM mechanism based on monotonic allocation and critical price theory to solve the problem of winning user decision-making and pricing. We demonstrate the economic characteristics of the mechanism, including truthfulness, individual rationality, and budget feasibility. The experimental results indicate the effective performance of the designed mechanisms.