One of the interesting topics in grid computing systems is resources discovery. After the failure of a resource in a chain of resources made for a specific task in grid environment, discovering and finding a new resou...One of the interesting topics in grid computing systems is resources discovery. After the failure of a resource in a chain of resources made for a specific task in grid environment, discovering and finding a new resource that reconstructs the chain is an important topic. In this study, with defining new agent that is called task agent, and by proposing an algorithm, we will increase the fault tolerance against probable failure of a resource in the resource chain.展开更多
The rapid expansion of enterprises makes product collaborative design （PCD） a critical issue under the distributed heterogeneous environment, but as the collaborative task of large-scale network becomes more complic...The rapid expansion of enterprises makes product collaborative design （PCD） a critical issue under the distributed heterogeneous environment, but as the collaborative task of large-scale network becomes more complicated, neither unified task decomposition and allocation methodology nor Agent-based network management platform can satisfy the increasing demands. In this paper, to meet requirements of PCD for distributed product development, a collaborative design mechanism based on the thought of modularity and the Agent technology is presented. First, the top-down 4-tier process model based on task-oriented modular and Agent is constructed for PCD after analyzing the mapping relationships between requirements and functions in the collaborative design. Second, on basis of sub-task decomposition for PCD based on a mixed method, the mathematic model of task-oriented modular based on multi-objective optimization is established to maximize the module cohesion degree and minimize the module coupling degree, while considering the module executable degree as a restriction. The mathematic model is optimized and simulated by the modified PSO, and the decomposed modules are obtained. Finally, the Agent structure model for collaborative design is put forward, and the optimism matching Agents are selected by using similarity algorithm to implement different task-modules by the integrated reasoning and decision-making mechanism with the behavioral model of collaborative design Agents. With the results of experimental studies for automobile collaborative design, the feasibility and efficiency of this methodology of task-oriented modular and Agent-based collaborative design in the distributed heterogeneous environment are verified. On this basis, an integrative automobile collaborative R＆D platform is developed. This research provides an effective platform for automobile manufacturing enterprises to achieve PCD, and helps to promote product numeralization collaborative R＆D and management development.展开更多
Traditionally, heuristic re-planning algorithms are used to tackle the problem of dynamic task planning for multiple satellites. However, the traditional heuristic strategies depend on the concrete tasks, which often ...Traditionally, heuristic re-planning algorithms are used to tackle the problem of dynamic task planning for multiple satellites. However, the traditional heuristic strategies depend on the concrete tasks, which often affect the result’s optimality. Noticing that the historical information of cooperative task planning will impact the latter planning results, we propose a hybrid learning algorithm for dynamic multi-satellite task planning, which is based on the multi-agent reinforcement learning of policy iteration and the transfer learning. The reinforcement learning strategy of each satellite is described with neural networks. The policy neural network individuals with the best topological structure and weights are found by applying co-evolutionary search iteratively. To avoid the failure of the historical learning caused by the randomly occurring observation requests, a novel approach is proposed to balance the quality and efficiency of the task planning, which converts the historical learning strategy to the current initial learning strategy by applying the transfer learning algorithm. The simulations and analysis show the feasibility and adaptability of the proposed approach especially for the situation with randomly occurring observation requests.展开更多
文摘One of the interesting topics in grid computing systems is resources discovery. After the failure of a resource in a chain of resources made for a specific task in grid environment, discovering and finding a new resource that reconstructs the chain is an important topic. In this study, with defining new agent that is called task agent, and by proposing an algorithm, we will increase the fault tolerance against probable failure of a resource in the resource chain.
基金Supported by National Science and Technology Major Project of China(Grant No.2009ZX04014-103)PhD Programs Foundation of Ministry of Education of China(Grant No.20100072110038)+1 种基金National Natural Science Foundation of China(Grant Nos.61075064,61034004,61005090)Program for New Century Excellent Talents in University of Ministry of Education of China(Grant No.NECT-10-0633)
文摘The rapid expansion of enterprises makes product collaborative design （PCD） a critical issue under the distributed heterogeneous environment, but as the collaborative task of large-scale network becomes more complicated, neither unified task decomposition and allocation methodology nor Agent-based network management platform can satisfy the increasing demands. In this paper, to meet requirements of PCD for distributed product development, a collaborative design mechanism based on the thought of modularity and the Agent technology is presented. First, the top-down 4-tier process model based on task-oriented modular and Agent is constructed for PCD after analyzing the mapping relationships between requirements and functions in the collaborative design. Second, on basis of sub-task decomposition for PCD based on a mixed method, the mathematic model of task-oriented modular based on multi-objective optimization is established to maximize the module cohesion degree and minimize the module coupling degree, while considering the module executable degree as a restriction. The mathematic model is optimized and simulated by the modified PSO, and the decomposed modules are obtained. Finally, the Agent structure model for collaborative design is put forward, and the optimism matching Agents are selected by using similarity algorithm to implement different task-modules by the integrated reasoning and decision-making mechanism with the behavioral model of collaborative design Agents. With the results of experimental studies for automobile collaborative design, the feasibility and efficiency of this methodology of task-oriented modular and Agent-based collaborative design in the distributed heterogeneous environment are verified. On this basis, an integrative automobile collaborative R＆D platform is developed. This research provides an effective platform for automobile manufacturing enterprises to achieve PCD, and helps to promote product numeralization collaborative R＆D and management development.
文摘Traditionally, heuristic re-planning algorithms are used to tackle the problem of dynamic task planning for multiple satellites. However, the traditional heuristic strategies depend on the concrete tasks, which often affect the result’s optimality. Noticing that the historical information of cooperative task planning will impact the latter planning results, we propose a hybrid learning algorithm for dynamic multi-satellite task planning, which is based on the multi-agent reinforcement learning of policy iteration and the transfer learning. The reinforcement learning strategy of each satellite is described with neural networks. The policy neural network individuals with the best topological structure and weights are found by applying co-evolutionary search iteratively. To avoid the failure of the historical learning caused by the randomly occurring observation requests, a novel approach is proposed to balance the quality and efficiency of the task planning, which converts the historical learning strategy to the current initial learning strategy by applying the transfer learning algorithm. The simulations and analysis show the feasibility and adaptability of the proposed approach especially for the situation with randomly occurring observation requests.
