A dynamic task offloading algorithm based on greedy matching in vehicle network

Abstract

With the rapid development of 5G technology, mobile applications such as autonomous driving, video streaming and vehicle online games are constantly emerging and data exchanges and service requests for portable terminal devices are increasing. The expeditious growth of data has brought a heavy burden to the roadside units (RSUs) and networks, the cellular networks cannot maintain the user's service quality. At present, vehicle fog computing is an effective method to solve the above problems in vehicle networks. However, the high mobility of vehicles and the complexity of traffic have brought great challenges to communication and computing in vehicle fog computing. To tackle the aforementioned problems, at first, a vehicle movement model VMM is presented to adapt the dynamics of vehicles in a traffic environment. The model uses four-lane dual carriageway to simulate the urban traffic environment. Next, in order to improve user's service quality (minimize the response time of the tasks), we propose KMM algorithm under the tasks information is known. It uses a two-time selection mechanism to select the offload server and uses the Kuhn-Munkras algorithm for the final decision. At last, the GMDC algorithm is proposed to adapt the dynamics of traffic environment. In this algorithm, the user vehicles appear randomly, we find a feasible offload server for it through 2 selections, and finally use the greedy algorithm to determine the most suitable server. Experimental results show that compared with the TOPM algorithm, our algorithm increases the task offloading rate by 5% and reduces the utilization rate of RSUs by 45% while improving the response time by 3%.