The European Commission emphasizes mass transit system resilience for a successful shift to sustainable mobility, urging specific actions. Urban resilience, emphasizing public transport performance amid various hazards, prioritizes redundancy through network restructuring to enhance disruption resilience. This study aims to design efficient transit networks, while also enhancing resilience against link failures by maximizing alternative paths for passengers. In the multi-objective setting considered, the goal is to obtain non-dominated solutions, considering two distinct objectives: maximizing path redundancy and minimizing passenger inconvenience, captured by a weighted score of average travel time and transfer shares. To obtain high-quality results, this study leverages reinforcement learning (RL) and multi-objective Particle Swarm Optimization (MOPSO), under the novel concept of intelligent optimization, where a learning component is used to guide the search. The proposed MOQLPSO exploits an adaptive search mechanism, with particles acting as self-interested agents within the solution space, incorporating dominance rules within the RL reward function. Benchmarking against a naive MOPSO version on two literature-based networks reveals the proposed approach's superiority. Results highlight trade-offs between efficiency and resilience, demonstrating denser and higher-quality pareto fronts in less computational time. The study underscores the importance of the reward function and emphasizes benchmarking as crucial for advancing the proposed concept.