Study Of Fog Node Deployment Optimization Problem For Industrial Io T Applications

As an emerging computing model,fog computing extends computing and storage capacity to the edge of the network,and can better meet time-sensitive data requests.Reasonable deployment of fog nodes is an important factor that affects the performance of fog network services,Therefore,it is necessary to study the deployment optimization problem of fog nodes.This thesis focuses on the fog node deployment optimization problem in typical industrial application scenarios.The aim to minimize the deployment costs under multiple constraints such as delay,throughput,and connection number constraints.The main content of this paper is concluded as follows.1)For the industrial scenario where some factories have a clear area division,a fog node discrete deployment problem model is proposed with multiple constraints of delay,number of connections,and bandwidth.A greedy and immune hybrid algorithm(GIHA)algorithm is then proposed,with the immune algorithm solving the fog node selection subset,the greedy algorithm solving the connection relationship between the sensor node and the fog node.The problem is decoupled into two subproblems,which reduces the complexity of the problem.The GIHA algorithm is proved that improves the search efficiency by the experimental results.Even when the network size becomes larger,highquality feasible solutions that satisfy all the constraints can always be obtained.2)For the industrial scenarios where the location of fog nodes in some factories is not limited,a continuous deployment problem model of fog nodes is proposed with multiple constraints of delay,number of connections,and bandwidth.An improved immune-based K-means algorithm(IIKA)is proposed.Weighted distance and balanced clustering are used to solve the constraints,and an immune algorithm is used to generate the initial solution to avoid the K-means algorithm's sensitivity to the initial clustering center.Experimental results show that compared with K-means algorithm and immune algorithm,the proposed IIKA algorithm can further optimize the deployment cost on the basis of ensuring the feasibility of the solution.3)For the industrial scenarios where the existing fog nodes cannot meet the increase in the number of sensor nodes,a problem model of incremental deployment of fog nodes is proposed.A heuristic algorithm based on density clustering is proposed.The sensor nodes divided into different clusters.Each cluster implement the deployment of sensor nodes and fog nodes applied with an improved Kmeans method.The algorithm solves the initial solution selection problem of the K-means algorithm through density clustering,and at the same time decomposes the complex problem into smaller subproblems,reducing the complexity of the problem.Experimental results show that,compared with the random deployment algorithm,the heuristic algorithm can effectively solve the problem of largescale incremental deployment of fog nodes,and has good scalability.