A novel container scheduling algorithm for edge computing: integrating entropy weight method and multi-strategy particle swarm optimization

A novel container scheduling algorithm for edge computing: integrating entropy weight method and multi-strategy particle swarm optimization

연구 분야: Networking

학회: Cluster Computing

With the rise of cloud and edge computing, container technology as a new virtualization means strongly supports the development of edge computing due to its deployment flexibility, excellent port-ability and scalability. However, edge nodes often face the challenge of load imbalance in actual operation, and excessive network communication load can also become a performance constraint. To address these issues, this paper proposes an innovative container scheduling algorithm that combines the Entropy Weight Method (EWM) and Multi Strategy Particle Swarm Optimization (MSPSO). By calculating the entropy of multiple objective functions and using a multi-strategy particle swarm algorithm to rationally assign weights. The algorithm effectively solves objective trade off problem and ensures efficient use of resources. To evaluate the performance of our proposed method, Random, Pack and BP-HP are used for comparison. The results show that compared with existing algorithms, EWM-MSPSO algorithm minimizes data exchange between servers while ensuring load balancing.