We consider a sharing economy over a network in which each vertex agent allocates resources to its neighbors in response to their contributions. General equilibrium theory can be applied here to solve the problem of deciding how to fairly and efficiently allocate resources among agents as resource sharing over the network can be modeled as a pure exchange economy. It is known that proportional sharing dynamics converges to a market equilibrium solution. We are particularly interested in proportional sharing dynamics as a mechanism for network resource sharing. Our focus is on the key issue in internet market design: whether an agent may manipulate its report of its own private information to gain more resources under this mechanism. This work establishes the first mathematical proof that such a practical distributed network resource-sharing protocol is truthful against the manipulative strategies of feasible weight misreporting and edge deletion applied both individually and together. Funding: This work was supported by the National Nature Science Foundation of China [Grants 11871366, 62172012]; Beijing Outstanding Young Scientist Program [Grant BJJWZYJH012019100020098]; the Fundamental Research Funds for the Central Universities; the Research Funds of Renmin University of China [Grant 297522503709]; and the Intelligent Social Governance Platform, Major Innovation & Planning Interdisciplinary Platform for the “Double-First Class” Initiative, Renmin University of China.

With mobile devices extensively used in daily life, there are ample opportunities to exchange sensing data through them, even without centralized management. In this paper, we design a peer based data exchanging model, where relay nodes move to certain locations to connect data providers and consumers to facilitate data delivery. Consumers are willing to pay for the data and these rewards are given to both relays and data providers. We first prove the NP-hardness of the problem on how to assign relay nodes to proper locations, and present a centralized optimal method with an approximation ratio. Then we define an autonomous compensation game for relays to make their individual decisions without any central authority. The sufficient and necessary condition for the existence of Nash equilibrium is derived, and an efficient reinforcement learning solver is designed to find the exact forms of equilibria. We analyze and compare this distributed game to the centralized social optimal solution, showing that the game incurs small bounded social costs, and is efficient under various network sizes, number of providers, number of consumers and device mobility.