<p>University Course Scheduling (UCS) is a very constrained, non-linear optimization problem that has a direct effect on institutional efficiency and student satisfaction. The present paper suggests an Intelligent University Timetabling framework, based on the combination of Particle Swarm Optimization (PSO) and Reinforcement Learning (RL) by means of Learning Automata (LA). In contrast to the conventional meta-heuristics, the suggested RLPSO framework employs a supervised LA model to track the success of the scheduling actions and to adaptively steer the population to the areas in the search space that are feasible. The model explicitly includes a set of hard constraints (professor availability, room capacity and equipment requirements) as well as soft constraints (reducing conflict of sessions between students). The experimental findings on five problem cases of different complexity indicate that the RLPSO offers significant performance improvements over traditional PSO and representative heuristics such as Graph Coloring and tsuGA on the evaluated problem instances. In particular, the suggested framework can reduce the hard constraint violations by up to 52% and soft conflicts by 88% in large-scale settings. In addition, the convergence speed increases by 41% with the incorporation of RL-based population guidance to arrive at optimal solutions in about 280 cycles compared to the 480 cycles taken by conventional PSO. These findings highlight the potential usefulness of the framework for practical university timetabling, though future validation on public competition-grade benchmark datasets is required to confirm its broad real-world applicability.</p>