Be the Best Sensors’ Manufacturer and One Stop Solution Service

Attn:Tara WhatsApp:+86 13951751584

a market based approach for sensor resource allocation in the grid.

by:KJTDQ     2020-05-12
This paper studies a market-based sensor resource allocation method.
This paper proposes an effective mechanism to assign sensor resources to the appropriate sensor grid users on the basis of the outcome of the participant negotiation.
The sensor allocation problem is modeled by introducing the sensor utility function.
Our goal is to find a sensor resource allocation that maximizes the total profit.
A distributed algorithm for optimal allocation of sensor resources is proposed.
The performance of the algorithm is evaluated and compared with other sensor grid resource allocation algorithms.
Key words: Sensors, grid computing, market Povzetek: Predstavljena ja postavitev senzorjev omrezje na osnovi pogajanj udelezenci med.
1 Introduction Grid computing is based on the concept of collaborative sharing of distributed and heterogeneous resources to solve the problem of large
Scale problem in dynamic virtual organization.
The grid computing model can be extended to include the sharing of sensor resources in the sensor network.
Fusion of sensor network and grid computing
Grid computing is like giving the computing grid \"eyes\" and \"ears. Real-
Time information about phenomena in the physical world can be processed, modeled, associated, and mined to allowthe-
Large-scale flight decision-making and operations [1, 16].
The sensor grid extends the grid computing mode to the sharing of sensor resources in wireless sensor networks.
The sensor grid combines the complementary advantages of sensor network and grid computing to support applications that require real-time computing.
Time information from the physical environment and a large amount of computing and storage resources.
Including environmental monitoring for the prediction and early warning of natural disasters, as well as the detection, tracking and interception of missiles [3].
However, the sensor-supported grid is an extensive distributed system that may consist of many sensors belonging to individual users who are not aware of the situation of other users at all.
They want to do what is best for themselves, which means they are rational and selfish.
Therefore, the sensor grid needs to provide incentives to encourage each user to contribute their resources to other users.
Sensor Grid is a relatively new research field, and there are still many problems that have not been solved in its design.
A major challenge in sensor-supported grid design is how to efficiently schedule sensor resources to user jobs in the collection of sensor resources.
Due to the limitation of energy, but also to extend the life of the sensor network, energy saving is an important consideration when managing the sensor network.
How to make better use of limited sensor resources to meet the needs of grid users without conflict and waste of resources is the key problem of sensor grid.
The resource allocation problem in grid computing system has been widely studied in the past.
There are some important differences between sensor resources and computing resources.
Therefore, in the grid supported by sensors, the existing traditional grid environment allocation algorithm may not work well.
A market-based sensor resource allocation method is proposed in this paper.
Since the task of the sensor user may compete for exclusive use of the same sensing resource, we need to assign a single sensor to the task of the sensor user.
The sensor grid task is characterized by uncertain requirements for perceived resource capabilities.
The sensor utility function is introduced to model the distribution problem.
Our goal is to find a sensor resource allocation that maximizes the total profit.
A distributed algorithm for optimal allocation of sensor resources is proposed.
The performance of the algorithm is evaluated and compared with other resource allocation algorithms in the sensor grid.
An application example of this method is given.
The rest of the paper is structured as follows.
The second section discusses related work.
In section 3, a market-based grid sensor resource allocation method is proposed.
The fourth section introduces the sensor resource allocation algorithm.
The experiment was conducted and discussed in Section V.
Section 6 gives the conclusion of the paper.
2 Related work for the problem of the combination of grid environment and wireless sensor network [1~17]
Incorporate sensors as consumers of grid resources into existing grid systems and provide sensor services to other grid nodes.
Peter Komisarczuk, etc. [1]
This paper discusses the research direction of malicious behavior detection, analysis and countermeasure in network sensor grid.
They outlined some of the experiences of these sensors and analyzed Network telescope data through grid computing as part of the \"smart layer\" of the Internet. M.
Pallikonda Rajasekaran, etc. [2]
A wireless sensor grid architecture is proposed for monitoring the health status of patients in different groups, providing a platform for doctors and researchers to share information with distributed databases and computing resources, to facilitate analysis, prognosis and drug delivery.
Hock Beng Lim and others. [3]
Design an integrated and flexible scheduler for the SPRING Framework-based sensor grid test platform.
A variety of scheduling and load balancing algorithms are implemented in this scheduler to adapt to the unique features of sensor jobs.
The scheduler can use appropriate scheduling or load balancing algorithms based on the needs of resource owners and users.
Nikolaos Preve, etc. [4]
The sensor grid enhanced data management system SEGEDMA is proposed, which guarantees the integration of different network technologies and the continuous data access of system users.
The main contribution is to realize the interoperation of these two technologies through a new network architecture. Huang-Chen Lee et al. [5]
Discuss the precautions for designing low-cost WSN-
It provides high resolution precipitation mapping.
Preliminary experimental results are given. Fox, G. et al. [6]
A collaborative sensor grid framework supporting the integration of sensor grid with collaborative grid and other grids is proposed.
The framework includes grid builder tools for discovering and managing grid services and remote distributed sensors.
It provides a real
The time collaboration client enables distributed stakeholders to view the displayed sensor stream consistently.
They demonstrate the versatility of the framework by building a robot-based customizable application to share situational awareness.
Hock Beng Lim and others. [7]
Designed to build a large
Seamlessly integrate the sensor grid infrastructure of heterogeneous sensor resources distributed in different projects in a wide geographic area. Sanabria, J. et al. [8]
A deployment framework is discussed, which utilizes existing grid computing technologies to provide middleware that integrates wireless sensor networks and grid infrastructure.
They demonstrate the work of enabling the sensor grid infrastructure for acoustic monitoring applications.
Hock Beng Lim and others. [9]
A framework of sensor grid architecture called extensible agent is developed.
Infrastructure of sensor grid (SPRING).
They designed the national meteorological sensor grid (NWSG), a large-scale cyber-
Sensor infrastructure for environmental monitoring.
NWSG integrates micro weather stations deployed across Singapore for weather data collection, processing and management.
Matsui and others. [10]
The distributed resource allocation problem of distributed sensor network is handled.
Use distributed optimization algorithms to understand problems and design collaboration protocols.
A distributed collaborative observation system using agent model is developed.
Xin Yujie and others. [11]
The architecture of wireless sensor grid is described, and a connection platform MPAS is designed.
The advantage of MPAS is that based on the Web service resource framework, it can combine multiple sensor networks with grids;
It can also start the sensor network and support the interoperation between multiple sensor networks.
Mohammed Mehdi Hassan and others. [12]
Discuss one of the most important issues in sensorsGrid, i. e.
, Develop fast and flexible content-
Information dissemination based on publication/subscription (CBPSID)
Systems that automatically fuse, interpret, share and deliver large amounts of sensor data to consumers as a whole sensor-
The grid environment is dynamic. Se-Jin Oh et al. [13]
Show a u-
Healthcare Sensor Grid gateway transparently connects sensor networks and grid networks to provide you with convenient and fast u-
Provide health care services to users.
They use mobile devices such as PDA to implement a mobile monitoring system to monitor the status of patients.
Li Xiaolin and others. [14]
Propose a framework for independent management (ASGrid)
Meet the needs of large emerging enterprises
Expand Applications in hybrid grid and sensor network systems.
They put forward the concept of autonomous sensor grid system in a holistic way.
Trivial large applications.
Matsui and others. [15]
A distributed resource allocation model for distributed sensor networks is proposed.
Several models based on constraint network and another model based on agent concept are compared.
A formal approach to constraint network for resource allocation problems similar to the proxy model is given. Yong-Kang Ji et al. [17]
The selfish problem of sensor network is discussed, and these problems are solved by using the mechanism of specific design, and several scenarios of sensor network application are described. The works [18~22]
It mainly deals with the problems of resource allocation and QoS optimization in the computing grid, and does not consider using sensor services to support sensor-supported grids.
[Difference of Paper]20][21]
This paper studies the optimal allocation of sensor resources. 20]
Two-level market solution for optimal resource scheduling in computational gridpaper [21]
Consider multiple QoS optimizations in the computing grid.
The computing grid is different from the sensor grid.
In the sensor grid environment, sensor nodes have serious limitations in terms of processing power, storage power, energy, etc. , but the computing grid does not have energy problems, but also has enough processing power.
Taking into account all these limitations of sensor nodes, it is very important that sensor grid systems manage energy consumption without affecting system performance.
The proposed sensor grid optimization considers the energy consumption of sensor nodes.
The objective of this paper is to maximize the utility of the system without exceeding the total amount of energy available, the cost budget and the deadline.
3 A method of sensor resource allocation in a market-based grid 3.
1 System model this paper uses the framework of computational economy to formulate the optimal allocation of sensor resources in the sensor-supported grid computing environment.
The model consists of two types of agents: the sensor resource agent representing the economic benefits of the sensor grid underlying sensor resource provider, and the sensor user agent representing the interests of the grid user uses the grid to achieve the goal.
The interaction between these two types of agents is regulated by market mechanisms.
The market mechanism in economics is based on distributed
The price change reflects the supply and demand relationship of resources, and the market theory in economics provides an accurate description of the efficiency of resource scheduling.
Sensor User agents can specify their requirements and preferences through utility models.
Therefore, the market-based sensor grid model essentially supports sensor users who have different requirements for sensor work execution.
The utility value is calculated by the provided utility function, which can be represented by the sensor working parameters.
The scheduler of the grid market analyzes the request.
Whenever a new sensor user agent is created, it is first given an electronic cash to complete the sensor work.
Sensor work can be described by deadlines, budgets, and sensor task requirements.
The sensor grid market mechanism allows multiple sensor resource agents and sensor user agents to negotiate at the same time, and it uses price-
Orientation method for allocating appropriate sensor resources. In this price-
Orientation Method to announce a set of initial prices to the sensor user agent.
Sensor users can update their allocation according to the price policy of the sensor provider and find the best solution iteratively.
In each iteration, the sensor user agent determines its optimal allocation separately and communicates its results to the sensor resource agent.
The sensor resource agent then updates their price and communicates the new price to the sensor user agent, and the loop repeats.
Prices then change over and over again to meet the demand for resources until the total demand is equal to the total amount of available sensor resources.
Sensor resource agents release sensor descriptions to the market.
Sensor suppliers actively compete for sensor work for sensor users and perform these work for profit.
Each sensor provider tries to maximize profits based on its resource capabilities and energy consumption.
We assume that the sensor resource agent does not cooperate.
On the contrary, their behavior is not
Cooperation aimed at maximizing personal profits.
The sensor resource agents compete with each other to serve the sensor user agents.
Sensor User agents also don\'t collaborate and try to purchase as many sensor resources as possible with the goal of maximizing their net benefits. 3.
2 in this section, we establish a mathematical model for optimal allocation of sensor resources in a sensor-supported grid computing environment.
First give the symbols used in the following sections :[p. sub. j]
: Price of sensor resource unit in sensor j[B. sub. i]
: Cost budget for sensor grid application I [E. sub. j]
: Energy limit of sensor resources j [u. sup. j. sub. i]
: Sensor Grid user I [money paid to sensor resource j]q. sup. n. sub. i]
: The sensor task of Ith sensor grid application n sensor jobs [t. sup. n. sub. i]
: Time required for the sensor grid application I to complete the sensor Job n [T. sub. i]
: Time limit given by sensor grid application I to complete all sensor jobs [es. sub. j]
Energy consumption per unit of sensor resources j [ec. sup. j. sub. i]
: Energy consumption of sensor resource j to complete sensor grid application I [x. sup. j. sub. i]
: The sensor resource j is assigned to the sensor of the sensor grid application I. It is assumed that the sensor grid system consists of multiple grid sites containing sensor nodes and ordinary fixed grid nodes.
The sensor node is composed of sensors connected by the sensor network.
In the sensor grid, an application set can be represented as an optimization for the allocation of sensor resources in the sensor grid, as shown below. [
Mathematical expressions that cannot be reproduced in ASCII. ](3. 2)[
Mathematical expressions that cannot be reproduced in ASCII. ](3. 3)Equation 3.
2 is the expression of the most effective sensor grid system.
The sensor grid utility is defined as the sum of utilities for all sensor users.
Costs incurred for the completion of the work indirect costs cannot exceed the cost budget [B. sub. i].
I can\'t exceed the deadline for completing all sensor work for the user application [T. sub. i].
Sensor j the total energy consumed by performing sensor user applications cannot exceed the energy limit [E. sub. j].
Total Allocated resources do not exceed total sensor capacity [c. sub. j].
We can solve such problems with the LA\'s method.
The problem of constraint optimization is solved by the method of drawing.
Let\'s consider the LA\'s form of sensor resource allocation optimization problem in sensor grid :[
Mathematical expressions that cannot be reproduced in ASCII. ](3. 4)[
Mathematical expressions that cannot be reproduced in ASCII. ](3. 5)[
Mathematical expressions that cannot be reproduced in ASCII. ](3. 6)Where [[lambda]. sub. i], [[beta]. sub. i], [gamma]. sub. i]
It is a la multiplier of sensor user I.
So let\'s assume that the sensor grid knows the utility function [U. sub. i]([x. sup. j. sub. i])
In all sensor user I, this optimization problem is mathematically manageable.
However, in practice, it is impossible to know all [U. sub. i]([x. sup. j. sub. i])
The centralized calculation and allocation of sensor resources in the sensor grid environment is also not feasible.
Max for objective function [U. sub. sensorGrid]
Global coordination of all sensor users is required, which is impractical in distributed environments such as sensor grids.
The model provided by the system (3. 2)
It is a nonlinear optimization problem with N decision variables.
Since Lagrangian is discrete, sensor users and sensor providers can handle the maximization of Lagrangian in parallel, respectively.
Sensor resource allocation {[x. sup. j. sub. i]}
Solve the problem (3. 2)
If and only if there is a set of non-negative shadow costs {[[gamma]. sub. i]}.
Usually, typical algorithms such as the fastest decent method and gradient projection method are used to solve such problems. The calculation complexity is high, the time is long, and it is unrealistic to implement.
In order to reduce the computational complexity, we decompose the sensor utility optimization problem (3. 2)
Two sub-problems for sensor users and sensor providers.
Shadow cost proposes a mechanism to allocate sensor resource optimization between sensor users and sensor grids.
We consider [the LA\'s][gamma]. sub. i]
The price charged as a sensor resource provider in the sensor market, equation (3. 7)
Describe the behavior of sensor consumers in the sensor market and the equation (3. 8)
Describes the strategy of the sensor supplier as a sensor supplier.
By breaking down Cohn-
In the sensor market, the role of sensor users and sensor suppliers is separate, and the problem of concentration is (3. 2)
Can be converted into distributed problems.
Sensor suppliers charge for sensor users.
The cost paid by sensor users to sensor providers is to solve the optimal cost of sensor resource allocation in the sensor market.
We break down the problem into two sub-problems (3. 7)
Optimization problems for sensor users and (3. 8)
This is the sensor provider optimization problem, in the case that the sensor provider does not need to know the utility function of a single sensor user, to seek a distributed solution.
Two maximizing sub-problems correspond to sensor user optimization problems such (3. 7)[
Mathematical expressions that cannot be reproduced in ASCII. ](3. 7)
Optimization problems with sensor providers, such (3. 8). [
Mathematical expressions that cannot be reproduced in ASCII. ](3. 8)In (3. 7)
For the sensor user optimization problem, under the deadline constraint, the sensor user provides the sensor provider with the only optimal payment to maximize the benefit of the sensor user. [
Mathematical expressions that cannot be reproduced in ASCII. ]
Represents the currency surplus of the sensor user, which is obtained by budget subtracting payments made to the sensor provider.
So ,(3. 7)
It is at some point to finish the work for sensor users as soon as possible and get more surplus. In (3. 8)
For the optimization problem of sensor providers, different sensor providers calculate the optimal allocation of sensor resources to maximize their own revenue and minimize the energy consumption to complete sensor grid applications.
We can choose any other form for utilities that increase with [increase]x. sup. j. sub. i].
But we chose the log function because as the total amount of allocated sensor resources increases from zero and then slowly increases, the revenue increases rapidly from zero.
In addition, the log function is convenient in parsing, increasing, strictly concave, and continuous.
The interests of sensor suppliers are affected by the payment, allocated resources and energy consumption of sensor users.
This means that with the increase in allocated sensor resources and the increase in payment, revenue also increases with the decrease in energy consumption.
The goal of the sensor supplier is to maximize [u. sup. j. sub. i]log [x. sup. j. sub. i]and minimize [
Mathematical expressions that cannot be reproduced in ASCII. ]
Under the constraints of their energy limit
Sensor providers cannot consume more energy [E. sub. j]
This is the energy cap provided by the sensor provider. [
Mathematical expressions that cannot be reproduced in ASCII. ]
Indicates the energy surplus of the sensor provider obtained by the energy limit minus the energy consumption.
Therefore, the optimization framework provides a distributed method for the sum utility maximization problem.
The sensor user layer problem adapts the sensor user\'s resource requirements according to the current sensor resource conditions, while the sensor Resource Layer intelligently allocates the sensors needed for the upper layer.
Interaction between layers is now controlled by using variables [[gamma]. sub. i]
, Which is the price that sensor providers charge from sensor users and coordinates sensor user needs and sensor resource supply.
For the sensor user optimization problem, under the deadline constraint, the sensor user provides the sensor provider with the only optimal payment to maximize the satisfaction of the sensor user. In equation (3. 7), Let [p. sub. j]
Indicates the unit price of the sensor j, making the pricing strategy, and the requirements of some sensor user agents cannot be processed on time.
Low budget sensor user agent does not have enough money to buy the sensor and cannot complete the work by the deadline;
This leads to low allocation efficiency.
When the load coefficient is 0.
5. The distribution efficiency of MSA is 27% higher than that of FCFS.
Compared with FCFS and SJN, the allocation efficiency of MSA is reduced more slowly when the load factor increases.
When the load coefficient is 0.
6. The distribution efficiency of LLF is reduced to 54%, and the distribution efficiency of MSA is reduced to 84%.
The allocation efficiency of MSA is better than that of SJN, LLF and FCFS.
Consider the success rate of execution from the results of Figure 1
2. When the load coefficient is 0.
5. The execution success rate of SJN is 28% lower than that of using MSA.
When the load factor increases, the execution success rate of SJN and FCFS decreases rapidly.
SJN and FCFS scheduling algorithms do not take into account the optimization of sensor resource providers and sensor users, it wants to minimize the running time of sensor jobs.
The success rate of MSA execution is higher than that of MSA.
When the load factor increases, the request of the sensor user agent can be accepted into the system due to the increase in the system burden, so, the sensor user agent has fewer requests that can be successfully executed before the deadline. Fig.
3 shows the energy consumption ratio under different load factors.
When the load factor increases, more requests need to be processed within one interval and the energy consumption ratio increases.
When the load coefficient increases LF = 0.
7, the energy consumption of MSA is 25% more than LF = 0. 4.
Under the same load factor (LF=0. 8)
The energy consumption of MSA is 16% lower than that of EDF.
The energy consumption ratio of MSA and EDF is lower than that of SJN, LLF and FCFS. Fig.
4 shows the effect of the load factor on the response time.
The smaller the LF, the lower the response time. When LF=0.
7, the response time of the MSA is 30% more than the response time of LF = 0. 2.
SJN provides the shortest response time in all algorithms.
In other algorithms, LLF has the longest response time.
The LF value is low, the system load is light, and the sensor price provided by the sensor resource agent is cheap;
Low-cost sensor user agents can choose cheap sensor resources to complete the work before the deadline, so the satisfaction of the sensor user agent is very high.
When the load of the system is heavy, the price of sensor resources is expensive;
Some sensor user agents need more time to complete the task. [
Figure 1 slightly][
Figure 2:[
Figure 3 slightly][
Figure 4 slightly]
6 conclusion this paper proposes a market-based sensor resource allocation in a sensor-supported grid computing environment.
Since the task of the sensor user may compete for exclusive use of the same sensing resource, we need to assign a single sensor to the task of the sensor user.
The sensor grid task is characterized by uncertain requirements for perceived resource capabilities.
The sensor utility function is introduced to model the distribution problem.
Our goal is to find a sensor resource allocation that maximizes the total profit.
A distributed algorithm for optimal allocation of sensor resources is proposed.
The performance of the algorithm is evaluated and compared with other sensor grid resource allocation algorithms.
In the future, we will consider moving our approach to a real grid platform to test its feasibility.
The author thanks the editors and anonymous reviewers for their helpful comments and suggestions.
This work was partially supported by the National Natural Science Foundation (NSF)under grant (No. 60970064,No. 61171075)
China National Key Basic Research Program (973 Program)under Grant No.
201cb302601, the national key laboratory open fund for software development environment under the grant (No. SKLSDE-2011KF-01)
New Century Talents program of Chinese universities (NCET08-0806)
China Fok Yingtong Education Foundation (Grant No. 121067).
Any opinions, findings and conclusions are the views of the author and do not necessarily reflect the views of the above-mentioned bodies.
Received: August 5, 2011 reference [1]
Internet sensor grid: Experience of passive and active instruments, progress of IFIP Information and Communication Technology, page132-145, 2010 [2]M.
Paliconda Rajasthan state
Radhakrishnan (P.
Subbaraj, sensor grid application for patient monitoring, next generation computer system, yb26 (4):569-575 [3]
Hock Beng Lim and Danny Lee are integrated flexible schedulers for sensor grids, UIC 2007, LNCS 4611, pp. 567-578, 2007. [4]
Nikolaos Preve, SEGEDMA: Sensor Grid enhanced data management system for medical computing, expert system with application3):2371-2380 [5]Huang-Chen Lee, Chun-Yu Lin, Chuan-Yu Cho, Chen-Lung Chan,Yao-Min Fang,Bing-Zhong Jean Li
Ta King, design considerations for sensor grids for monitoring precipitation in debrisflow-
Minutes of the Ninth International Conference on Information Processing of ACM/IEEE sensor networks, 2010, 362-363 [6]Fox, G. ; Ho, A. ; Rui Wang; Chu, E. ; Isaac Kwan;
A collaborative sensor grid framework
Cooperation in technology and systems in 2008. CTS 2008.
Page 2008 International Seminar. 29-38. [7]Hock Beng Lim; Iqbal, M. ; Wenqiang Wang; Yuxia Yao; A Large-Scale Service-
Meeting for sensor grid infrastructure, consumer communications and networking, 2009. CCNC 2009.
IEEE 6 TH, 2009, pp: 1-2 [8]Sanabria, J. ; Rivera, W. ; Rodriguez, D. ; Yunes, Y. ; Ross, G. ; Sandoval, C. ;
Sensor Grid Framework for 1 NC, IMS and IDC 2009 acoustic monitoring applications. NCM \'09.
Joint International Conference of the fifth session, P. 2009: 31-38 [9]
Hock Beng Lim, Mudasser Iqbal, Wang Wenqiang, Yao Yuxia, National Weather Sensor Grid: large-scale cyber-
International Journal of sensor networks environmental monitoring sensor infrastructure1-2): 19-36 [10]
Songjing Junbo and Songwei Hong, Journal of International Intelligent Information and database systems, U23 4 (4):307-321, [11]Yan YuJie; Wang Shu; Hao Zhao;
MPAS: connection platform for 2005 Asian wireless sensor network integration with grid and communication-
Pacific Conference, P. 2005. 1000-1004 [12]
Mohammed Mehdi Hassan, Biao song and Yui
Nam Huh, a dynamic fast event matching algorithm for content
Sensor-based publishing/subscription information publishing system
Grid, Journal of Supercomputing, 54 (3):330365 [13]Se-Jin Oh; Chae-Woo Lee; u-
Healthcare Sensor Grid gateway for connecting wireless sensor networks and grid networks, advanced communication technology, 2008. ICACT 2008.
The Tenth International Conference, Volume: P. 2008. 827-831 [14]Xiaolin Li; Xinxin Liu; Huanyu Zhao; Nanyan Jiang; Parashar, M. ;
Autonomous management and application of hybrid sensor grid system
2008. Computer communications and networks. ICCCN \'08.
Minutes of 2008 International Conference, Page: 1-6. [15]
Songwei Hongyi, formalization of distributed collaborative sensor resource allocation, KES-
AMSTA 2008, LNAI 4953, pp. 292-301, 2008 [16]Chen-
Sensor Grid: integrated sensor network and grid computing, CSI communication, 24 pages-Khong Tham and Rajkumar ya29, Vol. 29, No.
1. Indian Computer Society (CSI)
MUMBAI, India, July 200517]Yong-
Kang Ji, Zhang Yi, Xu Zhicheng and Wu Minyou, real resource allocation in selfish sensor network, MSN 2007, LNCS 4864, pp. 683-693, 2007. [18]
Proxy framework for supporting computing grids, Journal of Systems and software, 2017-11-70 (1-2): 177-187 [19]
Li Chunlin, Li Yanhong, QoS optimization in hierarchical computing grid, information science, 2007,177 (15):3038-3059 [20]
Li Chunlin, Li Liyuan, a distributed utility company
Optimal scheduling of computing grid resources based on two-level market solution, parallel computing3-4):332-351 [21]
Li Chunlin, Li Yanhong, multi-service QoS optimization strategy based on utility
Journal of standard scheduling, parallel and distributed computing on grid, xuan67 (2):142-153 [22]
Gao IM, Li Chunlin, Jin Hui and Li Lei uan, computer science department, Wuhan University of Technology, Wuhan 430063, ChinaR. China E-
Email: Yahoo spring 74com. Cn, shoptu @ public. wh. hb.
Custom message
Chat Online 编辑模式下无法使用
Leave Your Message inputting...