Q-learning is an effective model-free reinforcement learning algorithm. However, Q-learning is centralized and competent only for single agent learning but not multi-agent learning because in later case the size of st...
详细信息
Q-learning is an effective model-free reinforcement learning algorithm. However, Q-learning is centralized and competent only for single agent learning but not multi-agent learning because in later case the size of state-action space is huge and will grow exponentially with the number of agents increasing. In the paper we present a distributed Q-learning algorithm to solving this problem. In our algorithm, the tasks of learning optimal action policy are distributed to each agent in team but not a central agent. In order to reduce the size of action-state space of multi-agent team we introduce a state-action space sharing strategy of agent team, through which one agent in team can use the states already explored by other agents before and need not take time to explore these states again. Additionally, our algorithm has the ability to allocate sub-goals dynamically among agents according to environment changing, which can make agent team coordinate more efficiently. Experiments show the efficiency of our algorithm when it is applied to the benchmark problem of predator-prey pursuit game, also called pursuit game, in which a team of predators coordinate to capture a prey.
A method for simultaneous analysis of the two components of compound paracetamol and diphenhydramine hydrochloride powdered drugs on near-infrared (NIR) spectroscopy is developed by using a Radial Basis Function (RBF)...
详细信息
Most of real-life scheduling problems are semi on-line scheduling,where we do not know all the time variables of jobs in ***,solving such dynamic problems is a hot topic of research in Constraint *** this paper we ana...
详细信息
Most of real-life scheduling problems are semi on-line scheduling,where we do not know all the time variables of jobs in ***,solving such dynamic problems is a hot topic of research in Constraint *** this paper we analyze the constraints in the semi on-line scheduling problems and introduce the concept of monotonic about constraints which appear in general dynamic constraint *** designed a generic dynamic constraint solving algorithm for solving the scheduling *** give an example of its application to semi on-line discrete resource-constrained scheduling problem(ship-loading problem).
In UML, a modeling language popularly used to describe models at different abstraction levels, sequence diagram and statechart are both used to represent systems' behavior models. A consistency problem may arise d...
详细信息
In UML, a modeling language popularly used to describe models at different abstraction levels, sequence diagram and statechart are both used to represent systems' behavior models. A consistency problem may arise due to the fact that sequence diagram represent interactive behavior of all objects, whereas statechart represent the complete behavior of one specific object. This paper proposes an approach checking the consistency between sequence diagrams and state diagrams. In our approach, statechart is formalized using finite state processes and sequence diagram is represented using messages trace. An existing tool LTSA is used to support our method.
This paper presents a method of medicine composition concentration analysis based on least square support vector machines (LS-SVMs) and examines the importance of the hyperparameter choice in improvement of algorithm ...
详细信息
This paper presents a method of medicine composition concentration analysis based on least square support vector machines (LS-SVMs) and examines the importance of the hyperparameter choice in improvement of algorithm performance. Simulation results show that the proposed method obtains high quality precision in the generalization, compared with multiple linear regression, and that it is an efficient approach to regression estimation.
The workflow model is the abstract expression of the workflow or the business process. Following the WfMC reference model, a PKI-based lightweight workflow model named as PBLW is put forward in this paper. The framewo...
详细信息
The workflow model is the abstract expression of the workflow or the business process. Following the WfMC reference model, a PKI-based lightweight workflow model named as PBLW is put forward in this paper. The framework of PBLW is proposed. Based on the time constraint directed graph, the workflow theory model in formalization is described in detail, and the time constraint problems are discussed in order to enhance the time management ability of the workflow model. Furthermore, in order to express the complicated secure access control constraint of the workflow, a conditioned RBAC model based on PKJ technology is presented, according to the traditional RBAC security access control model. Finally, the workflow engine driven algorithms for the workflow instance are given. By the actual inspection and users' feedback, it proves that this model can express the basic requirements of a majority of the process applications.
This paper presents two parallel semantics of constraint logic programs: multiset answer constraint semantics and game semantics, which differ entirely from the traditional semantics. When giving the first semantics, ...
详细信息
This paper presents two parallel semantics of constraint logic programs: multiset answer constraint semantics and game semantics, which differ entirely from the traditional semantics. When giving the first semantics, it adopts the s-semantics approach, whose aim is to find notions of models which really capture the operational semantics. And it defines the program equivalence based on multiset answer constraint. Then it investigates the game semantics and proves that it is consistent with traditional operational semantics. On the basis of these two semantics, programmers can comprehend the programs deeply and develop more sound parallel constraint logic programming systems. Finally, it gives the conclusions and future work.
ETL tools are responsible for the extraction of data from sources, their cleansing and loading into a target data warehouse. However, nowadays, the design and development of ETL processes are performed in an in-house ...
详细信息
ETL tools are responsible for the extraction of data from sources, their cleansing and loading into a target data warehouse. However, nowadays, the design and development of ETL processes are performed in an in-house fashion, and need uniformed methodological foundations. In this paper, we propose a novel conceptual model for the modeling of ETL processes. We employ CommonCubes to represent the cubes in a target data warehouse. CommonCubes release the design of ETL processes from overdependence on the physical schema of the target data warehouse, and enable the designers to pay more efforts to data transforming than data loading when designing ETL processes. Based on the constraint functions on source attributes and the transforming operations on target attributes, we define ETL mappings 1:0 capture the semantics of various relationship cardinalities between source attributes and target attributes, which provide a good basis for the design of ETL processes.
This paper presents a new model to incorporate decision theory into Graphplan framework, which enables our planner to handle uncertainty and make decision to choose the optimal one among a set of hypothesis valid plan...
详细信息
This paper presents a new model to incorporate decision theory into Graphplan framework, which enables our planner to handle uncertainty and make decision to choose the optimal one among a set of hypothesis valid plans. This planer, called UTDP is tested on several experimental domains. And the experimental results show that UTGP is sound and efficient and performs better than the famous probabilistic planner Buridan.
暂无评论