<p>In this paper, we propose a resource-oriented improved aquila optimization framework (RO-IAO) for the joint optimization of vehicular task offloading and resource allocation. Roadside units (RSUs) act as communication nodes and are connected to the remote cloud server via wired backhaul, with built-in mobile edge computing (MEC) servers, they enable task offloading and execution. Specifically, we first build a NOMA-MEC assisted internet of vehicle (IoV) system model with multiple vehicles and multiple RSUs, and formulate an objective that jointly optimizes latency and energy. Second, for resource allocation, we design a resource orientation strategy driven by a priority factor. Tasks are classified by urgency into type A tasks and type B tasks, weights are assigned based on channel variations and MEC load, and combined with the weights to allocate the system resources toward higher urgency type A tasks. Finally, the improved aquila optimization (IAO) algorithm is employed for IoV task offloading. IAO uses a tent chaotic map and reverse learning. It initializes the population by uniformly covering the feasible region with a tent sequence and generates opposition solutions when the search stalls, accelerating convergence and enhancing population diversity to cope with the highly time-varying and dynamic IoV setting. Extensive simulations show that the proposed scheme reduces the system average cost by 6.5%-36.7% compared with the baselines.</p>