Adaptive routing protocols for determining optimal paths in AI multi-agent systems: a priority- and learning-enhanced approach

TL;DR

This work tackles adaptive routing in AI multi-agent systems by extending Dijkstra's algorithm with a multi-factor, priority-aware cost function and reinforcement-learning (RL) driven weight adaptation. The Enhanced Adaptive Priority-Based Dijkstra's Algorithm (APBDA) incorporates factors such as task complexity $T$, priority $P$, agent capability $C_j$, bandwidth $B_{ij}$, latency $L_{ij}$, model sophistication $M_j$, load $F_j$, and reliability $R_j$, with weights $w_1$–$w_7$ learned over time to optimize both latency for critical tasks and overall resource utilization. To scale to large MAS, the approach adds heuristic filtering and a hierarchical routing structure, enabling intra- and inter-cluster routing with aggregated metrics. The RL framework updates the weight vector based on network-wide state (e.g., average latency, load distribution) and rewards routing decisions that minimize high-priority task completion time while balancing load and reliability. The proposed method promises context-aware, robust, and scalable routing suitable for federated learning platforms, distributed robotics, and large-scale AI deployments where dynamic conditions prevail.

Abstract

As distributed artificial intelligence (AI) and multi-agent architectures grow increasingly complex, the need for adaptive, context-aware routing becomes paramount. This paper introduces an enhanced, adaptive routing algorithm tailored for AI multi-agent networks, integrating priority-based cost functions and dynamic learning mechanisms. Building on an extended Dijkstra-based framework, we incorporate multi-faceted parameters such as task complexity, user request priority, agent capabilities, bandwidth, latency, load, model sophistication, and reliability. We further propose dynamically adaptive weighting factors, tuned via reinforcement learning (RL), to continuously evolve routing policies based on observed network performance. Additionally, heuristic filtering and hierarchical routing structures improve scalability and responsiveness. Our approach yields context-sensitive, load-aware, and priority-focused routing decisions that not only reduce latency for critical tasks but also optimize overall resource utilization, ultimately enhancing the robustness, flexibility, and efficiency of multi-agent systems.