Weighted Capacitated Popular Matching for Task Assignment in Multi-Camera Networks

Multi-Camera Networks (MCN) are becoming increasingly important in today's society needs and daily-life with application-oriented multiple tasks running in each camera such as video surveillance, object tracking and localization etc. The ultimate goal of MCN is to best satisfy such tasks' preferences/expectations required by users, which has not been well-addressed by previous works. This paper investigates such challenge by formulating a novel weighted capacitated Popular Matching for multi-Task assignments (PMT) problem and proposing efficient algorithms to solve the problem. Using the popularity to represent the optimality of task-camera matching, we can find a matching in which the allocation of the most tasks to the corresponding cameras is closest to the tasks' preferences. With extensive simulations, we demonstrate that our approaches can make matching to the satisfaction of all tasks efficiently as compared to those baseline approaches.