Resource-and-Latency-Aware PSO — A novel algorithm for intelligent and contextual cloud native resource scheduling

Swarm Intelligence (SI) has emerged as a prominent approach in resource scheduling, offering adaptabilityand effectiveness in multi-objective optimisation. However, SI-based methods often introduce complexity and challenges suchas premature convergence and excessive exploration, leadingsuboptimal performance. To address these challenges, this paperpresents a novel Resource-and-Latency-Aware Particle SwarmOptimisation (RALA-PSO) algorithm for intelligent and contextual cloud native resource scheduling. RALA-PSO expands uponParticle Swarm Optimisation (PSO) by incorporating a greedyinitialisation for faster convergence, a circular search spaceto avoid boundary stagnation, a multi-criteria fitness functiontailored to Cloud-Native characteristics, a heuristic-based scoringmodel to reward suitable solutions, and the integration of adaptive weights and learning factors for exploration–exploitationbalance. The RALA-PSO implementation is available as opensource. RALA-PSO prioritise solutions based on efficientresource allocation and less communication overhead via serviceco-location. Experimental evaluations using the Alibaba ClusterTraceV2018 dataset show that RALA-PSO reduces networktransmission overhead cost by 1.8× to 2.36× compared to state-of-the-art algorithms MOPPSO-CMS and ACO-CMS, whilemaintaining competitive resource efficiency. Notably, RALA-PSOachieves runtimes up to 212 times faster than ACO-CMS and 1.5times faster than MOPPSO-CMS. These results demonstrate thatRALA-PSO is a scalable, efficient, and context-aware schedulingsolution for cloud-native applications, particularly effective inenvironments with strict network and resource constraints.