Study of distributed Lagrangian heuristics for self-adaptive publish/subscribe network design problems

The Internet of Things (IoT) has revolutionized information collection and processing through the interconnection of smart objects that can transmit data for analysis. However, IoT devices typically send data to cloud servers, which can lead to connectivity and data transfer issues. Edge computing-based solutions are being studied as a solution, which involves processing data directly at the source to enable more efficient and effective services. However, current IoT infrastructures are not yet ready for this transition. One solution being explored is the use of multiple distributed MQTT brokers on different interconnected machines to improve system reliability and scalability. A fully-distributed optimization solution based on a Lagrangian relaxation approach is being considered to ensure optimal load balancing and reliability for the entire system. The objective is to evaluate the effectiveness of distributed Lagrangian heuristic algorithms in the field of communication network management, which allows network nodes to act autonomously based on information about themselves and neighboring nodes they can communicate with, without centralized management.