A resource allocation scheme based on Semi-Markov Decision Process for dynamic vehicular clouds

Abstract

Currently, the number of resources within a vehicle is growing. The vehicle can provide its idle resources to other vehicles through a cloud. Thus, these vehicles can communicate with each other to dynamically create a vehicular cloud. Therefore, this vehicular cloud needs to adapt according to the number of available resources that the vehicle members in the cloud are sharing. In this kind of cloud, the control of allocated and shared resources becomes a challenge due to the high mobility of vehicles. With this challenge in mind, we propose an optimal resource allocation scheme in order to maximize the use of the available resources. The optimal problem to maximize the expected average reward system is formulated as a Semi-Markov Decision Process (SMDP). The SMDP problem is solved by an iterative algorithm. Numerical results have shown that the proposed scheme has a stable behavior independent of the frequency of requests or the amount of resources. Furthermore, the proposed solution keeps the block rate at 20%, priorizing the allocation that will maximize the utilization of the available resources.