A Stable Task Assignment Mechanism for Multi-Platform Mobile Crowdsensing

Abstract

Mobile crowdsensing (MCS) is a new paradigm for Internet of Things. It can fully utilize the smart devices carried by mobile users for accomplishing various sensing tasks. Most existing task assignment mechanisms have assumed the availability of just a single service platform and without considering the preference of users and platforms. In this paper, we focus on studying how to design efficient task assignment mechanism when there are multiple service platforms in the MCS system and further the preferences of both users and platforms are considered while the sensing quality of each user is unknown in advance. The design objective is to maximize the overall sensing qualities of all completed tasks while respecting the budget constraint of each platform. We build a multi-platform oriented task assignment framework and formulate the problem under study as a 0-1 integer linear programming (ILP) problem. We propose a Multi-platform Stable Task Assignment mechanism (MSTA). MSTA works in a round by round manner. In each round, MSTA first performs budget splitting among different task locations for each platform, then makes stable matching between users and platforms and performs online learning of users' sensing qualities by using the multi-armed bandit (MAB) model. We deduce the time complexity of MSTA and prove that MSTA has the properties of stability, individual rationality, budget feasibility, and truthfulness. Simulation results demonstrate the high performance of the proposed MSTA mechanism.