Research On Low Energy Consumption Computing Method For Drug Control Cloud Platform

Informatization is an important means to enhance the efficiency of drug control,which directly bears on the national economy and people's livelihood.The existing drug control systems are mainly based on single area business management software.It has been difficult to meet the challenges brought by users' explosive growth in drug testing information and the corresponding efficiency of authentication,access,acquisition and processing.Considering the status quo and trend of drug control informatization at home and abroad,this thesis clarifies that Drug Control Cloud Platform(DCCP)is the key to drug control informatization.This thesis is based on the drug control work flow of the task requirement group of drug control clients(here in after referred to as the task requirement group).We discussed the composition of DDCP,designed the logical structure,network architecture,system composition and application functions of DCCP,pointing out that energy consumption is essential to DCCP development.Then,the DCCP classification fusion energy consumption model was constructed,including three parts authentication and access,drug control information acquisition and task scheduling processing,and an energy-saving algorithm According to the characteristics of each part,the corresponding energy saving algorithm is proposed,and experiments are carried out through real energy consumption and network simulation.The results demonstrate that the proposed algorithm can reduce energy consumption without sacrificing the network performance.The research results in this thesis are of great significance for promoting DCCP to solve the problem of drug inspection information.The innovations of this paper are as follows:(1)The authentication and access control over the task requirement group prior to the access to the drug control cloud platform(DCCP)was analyzed;revealing that the energy consumption of the control process is related to client attributes,access behaviors and type of access.On this basis,three energy-saving and efficient security algorithms were put forward to save energy and satisfy the need of authentication and access control,including a Client Status Stable Evaluation(CSSE)for the authentication problem with constant user attributes,a Client Status Stable Access Process Strategy(CSSAPS)for the authentication problem with constant access behaviours,and a Client State and Access Process Change Evaluation(CSAPCE)for the authentication problem with variable client attributes and access behaviours.The CSSE was developed through classification and the definition of standard client status set,the CSSAPS was designed with hierarchical structure model and a security policy library,while the CSAPCE was created from client Dynamic Authentication and Access Process based on Markov Process(DAAP-MP).Through actual energy consumption experiments,it is proved that the CSSE consumed about 5%~23% less energy than the existing DCCP authentication algorithms,the CSSAPS consumed about 3%~26% less energy than the DCCP algorithms with security measures,and the CSAPCE saved lots of energy by re-authenticating the clients and relocating the accesses after the change in client attributes and access behaviours.In addition,network simulation shows that the three proposed algorithms performed well in the identification of illegal clients,the mean delay and network jitter,indicating that these algorithms can ensure the security and stability of the DCCP in static authentication,static access,dynamic authentication and access.(2)To reduce the energy consumption in the acquisition of DCCP drug control information,a Multi-Layer Protocol Improved Algorithm based on Neural Computing and Particle Swarm Optimization(PSO-NC-MLPIA)was proposed,and adopted to extract the main factors affecting the energy consumption in the acquisition of DCCP drug control information according to the demand,application scenario and experience of the task requirement group.These factors were taken as important parameters in the hidden layer of the neural computation process.After that,iterative neural computations were carried out based on these parameters and the real-time data acquired by data collection nodes.The results of the output layer make it possible to select the most suitable MLPIA for the drug information acquisition network,thereby reducing the energy consumption and improving the acquisition efficiency.Targeting different layers of the drug information acquisition network,the MLPIA at once optimizes the protocol algorithm for the medium access control(MAC)layer,improves the protocol algorithm for the routing layer,and introduces the algorithms for the application layer and the transmission layer.The PSO-NC-MLPIA was further optimized by the particle swarm optimization(PSO),aiming to avoid the local optimum trap and accelerate the search for the optimal solution.Through actual energy consumption experiments,it is proved that the PSO-NC-MLPIA consumed about 13%~37% less energy than protocol selection algorithms improved on a single network layer while fulfilling the needs of the same task group.In addition,network simulation verifies the good performance of the PSO-NC-MLPIA in packet loss rate,network life cycle,3D energy consumption curve and network delay.(3)The author presented a Task Scheduling and Processing Algorithm based on Genetic Algorithm and Tabu Search(TSPAGA-TS),considering the energy consumed to schedule the tasks of the task requirement group on the DCCP.Specifically,the resource allocation constraints on the task scheduling and processing were identified through the analysis on the multiple tasks,demands and conditions of the task requirement group.On this basis,the genetic coding method,fitness function,selection operator,crossover operator and mutation operator were designed for the tasks.Then,the tabu search was introduced to overcome the low efficiency in the search for the optimal solution.The crossover operator and mutation operator of the tabu search,which are more targeted to our problem than the original operators,accelerated the search for the optimal solution by increasing the population diversity.Through actual energy consumption experiments,it is proved that the TSPAGA-TS consumed about 5%~15% less energy than ordinary cloud platform resource scheduling algorithms and less time for the task requirement group to complete its tasks.In addition,network simulation verifies that the TSPAGA-TS can improve the resource utilization rate in task allocation and processing of the DCCP and maintain an excellent maximum response time.