<p>The rapid increase in renewable energy resource integration has resulted in significant generation uncertainty consequently causing supply demand imbalance. To address this challenge, electric vehicles (EVs) are increasingly recognized as a viable solution for providing load flexibility to distribution system operators (DSOs). However, incorporating multiple EVs into system optimization creates a high-dimensional problem, increasing complexity and computational burden. Thus, aggregation is necessary to reduce the number of constraints and simplify the optimization process. In response to the challenges of EV aggregation and their integration at system level, this paper proposes a distribution locational marginal price based bilevel optimization model, considering uncertainties in solar generation and interaction between DSO and EV aggregator. The objective of the model is to minimize the total operational cost of DSO and EVA payments in the worst situation of uncertainties. To account for the uncertainties associated to renewable power, a robust optimization approach is adopted. The bilevel model is transformed to single level using Karush- Kuhn- Tucker conditions and strong duality theorem. The case study illustrates the effectiveness of the proposed framework in addressing the impact of renewable power forecasting errors. The results show that load flexibility provided by aggregated EVs enhances the balancing and security of the distribution grid, while accommodating uncertainties in renewable energy generation. This research provides valuable insights into how DSO can improve its decision-making processes and ensure the stable operation of system in the face of increasing renewable integration.</p>