Cost Optimization And Design Research For Industrial CPS Application

With the development of Cyber-Physical Systems(CPS),the single and fixed production mode of traditional industrial automatic control systems has been changed,which has promoted the application of CPS in industry rapidly.That Embedded System is constituted of sensors,actuators,processors and heterogeneous networks can support complex,mass industrial production lines through broad and flexible ways.However,the issues of reliability and cost increased gradually caused by highly integrated systems of industrial CPS and large-scale production lines.Cost reduction and reliability guarantee are the important indexes for the application design and production of industrial CPS,and balancing the design and optimization of these system attributes is a great concern problem.This paper focuses on the cost optimization and design of Industrial CPS,which includes the cost optimization of application development for ACPS(Automotive Cyber-Physical Systems),redundancy optimization for CPCS(Cyber-Physical Cloud Systems)application and the cost design for CPCS applications based on limited budget.The main work and innovations of this paper are summarized as follows:1.The problem of resource cost optimization for reliable industrial CPS applications is studied.There are two issues of this study.One is the satisfaction of reliability index,which can be solved by dividing the whole application reliability index into subindex of each task,and the overall goal is achieved through satisfying the reliability of each scheduled task.The other is the optimization of resource consumption cost.This issue is solved by assigning each task to a processor with minimum resource consumption by a heuristic way,meanwhile the reliability of each task is satisfied.The experimental result in real industrial CPS applications environment shows that the MRCRG(Minimizing Resource Consumption Cost with Reliability Goal)algorithm proposed in this paper achieves the minimal resource consumption cost compared to the classical MaxRe and RR optimization algorithms.2.Three hardware cost optimization algorithms are proposed for reliable industrial CPS applications in the design phase: Exploratory Hardware Cost Optimization algorithm(EHCO),Enhanced EHCO algorithm(EEHCO)and Simplified EHECO algorithm(SEEHCO).EHCO reduces the hardware cost of the system through iterative removal of the processor while satisfying the functional safety of the application.EHCO can meet the real-time requirements of the application in some cases,but the reliability cannot be satisfied and the reliability enhancement algorithm RE is provided.EEHCO algorithm is proposed base on RE which enhance the reliability of the application to solve the mentioned problem.Finally,based on the consideration of the time complexity of EEHCO,a simplified SEEHCO algorithm is also proposed to solve the problem of minimizing the hardware cost of large-scale Industrial CPS applications.Experiments show that the lowest hardware cost can be obtained and compared to the IHCO,EEHCO algorithm generates the minimum hardware cost and SEEHCO is more effective for large-scale Industrial CPS applications than existing algorithms.3.The issue of development cost optimization for reliable ACPS applications is solved.This issue is decomposed into two sub-problems: calculating the reliability of the ASIL decomposition scheme and minimizing the development cost.For the decomposition scheme of the automotive safety integrity level ASIL provided by the ISO26262 standard,the RCS algorithm is proposed to calculate the reliability value of each ASIL decomposition scheme to solve the first sub-problem.The second subproblem is solved by proposing the Minimizing Development Cost with Reliability Goals algorithm(MDCRG)to select a decomposition scheme that can generate the minimum development cost under the constraint of the reliability of each scheme.The study is proved by experiments that the MDCRG algorithm can generate less development cost than the MDC and GARG algorithms and has lower time complexity than GARG.4.The problem of redundancy optimization for reliable CPCS applications is studied.The fault-tolerant techniques are used to achieve system reliability goals,and two redundancy minimization algorithms are proposed: Enough Replication for Redundancy Minimization algorithm(ERRM)and Heuristic Replication for Redundancy Minimization algorithm(HRRM).The ERRM algorithm selects the redundant task that generates the maximum reliability by iterative pre-selection based on the Lower Bound Reliability algorithm(LBR)until the reliability of the whole application is satisfied.Considering the high time complexity of the ERRM algorithm,the HRRM algorithm uses a heuristic method to improve the unbalance distribution of the reliability for the nearby of RR algorithm's entry task through the pre-allocation scheme to achieve the low time complexity and minimize redundancy.In the real CPCS applications environment generated at different scales,parallelism and heterogeneity,the results demonstrate that ERRM can produce minimal redundancy compared to MaxRe and RR algorithms while HRRM achieves approximate minimum redundancy with lower computation time.5.A Minimization Energy Consumption using Available Budget Pre-assignment algorithm(MECABP)is proposed to solve the problem of cost design with budget constrained for CPCS applications.This problem is decomposed into two sub-problems:transferring cost budget and minimizing energy consumption.The available budget pre-allocation method is provided using pre-allocation technology,and the cost budget problem of application is transferred to the task cost budget problem in order to solve the first sub-problem.Further more,MECABP is proposed to realize the energy consumption reduction under the constraint of meeting the task cost budget.Finally,through three DAG applications with different parallelism,it is proved that the proposed algorithm is more efficient than MECBL algorithm and EHBCS algorithm,and can generate the lowest energy consumption.