Pursuit-Evasion Game Model-Based Mobile Edge Computing System for Efficient Task Scheduling in a Dynamic Environment

Vehicle-to-infrastructure (V2I) connectivity systems are challenged further in this age when the number of vehicles seeking services from road side units (RSUs) is always increasing. This leads to difficulties in service provision and the allocation of resources. In this study, attempts are made to address the issue of connectivity and resource allocation in a very congested and highly trafficked area where existing approaches are limited in managing the variability in use and avoiding network overuse. We consider onboard units (OBUs) capable of such tasks and propose a time-varying attractive traffic circulation model that uses geolocated vehicles as a surrogate for trust. Furthermore, this allows OBUs to task-compute heavy workloads to the optimal RSU, thereby improving processing efficiency in terms of time and reducing computing costs. To account for the competitive delay, we handle extended execution time that captures data arbitrations and channel contesting in the Internet of vehicles (IoV) systems having a dynamic structure. In our research, to solve the problems posed by the existing methods, we propose a new task-scheduling strategy using the Pursuit-Evasion Game Model, which facilitates autonomous decision-making, improves scheduling versatility, and decreases processing delays. Based on a Genetic algorithm (GA), our Single Pursuer- Single Evader (SP-SE) scheduling mechanism determines the best available RSUs for task broadcasting, dynamically varies computation durations, and resolves channel contention. The simulation results verify the application of our method and illustrate that SpPs in adverse conditions perform better than solutions designed using classical GA-based scheduling.