<p>Software-Defined Networking (SDN) is an innovative network architecture that separates the control and data planes, enabling centralized network management through controllers in the control layer. However, dependence on a single controller may lead to network congestion, bottlenecks, or complete system failure in the event of a malfunction. To mitigate these issues, deploying multiple controllers is essential, with their number and placement determined by network size and traffic demands. This challenge is formally known as the Controller Placement Problem (CPP), which plays a key role in achieving efficient network performance in terms of delay, load balancing, reliability, and energy efficiency. To address the CPP, a hybrid approach is proposed that utilizes reinforcement learning algorithms to determine the optimal number of controllers, and heuristic algorithms for controller placement and backup path identification. The approach aims to optimize network delay, minimize energy consumption, and balance load distribution while maximizing the diversity of routing paths between controllers and switches. Chaos theory is employed to improve the diversity of the initial population, and solution discretization techniques are applied to align with the discrete nature of the problem. A novel metaheuristic algorithm, the Coati and Horse Optimization Algorithm (CHOA), is introduced by combining the Coati Optimization Algorithm with the Horse Herd Algorithm. Genetic operators are integrated to enhance solution refinement and discretization in each iteration. The performance of CHOA is evaluated on four widely used SDN topologies from the Internet Topology Zoo and compared against existing state-of-the-art algorithms. Results indicate that CHOA reduces load imbalance by up to 18%, propagation delay by up to 14%, and average energy consumption by up to 19%. Additionally, considering alternative paths between controllers and switches reduces the risk of disconnection in the event of a link failure by 26%.</p>