Adaptive container scheduling strategy for multidimensional perception

In edge computing environments, such as Kubernetes clusters, which often face challenges such as resource heterogeneity, network constraints, and high concurrency of tasks, traditional scheduling strategies can hardly meet the demands of multi-dimensional resource awareness and dynamic adaptation. To this end, an adaptive container scheduling algorithm, MAPSO, is proposed, which takes particle swarm optimization as the core and introduces chaotic initialization, dynamic inertia weights, and adaptive learning factors to improve the global search capability and convergence efficiency. At the same time, MAPSO obtains CPU, memory, disk I/O and network bandwidth status in real time through a multi-dimensional resource awareness mechanism, and constructs an adaptation function to guide the particles to search intelligently. MAPSO has been integrated into the Kubernetes scheduling framework, and verified in edge clusters. Experimental results show that MAPSO outperforms the default scheduler in terms of resource utilization, load balancing and scheduling stability, especially in I/O-type scenarios, and shows higher efficiency and adaptability, which has good engineering practical value.