Catálogo de publicaciones - libros

With the new software architecture appearing, the application performance evaluation becomes quite difficult. Based on previous studies, a new architecture is proposed concerning using multi-agent network data in the application performance evaluation. Architecture defined both active and passive methods to simulate and test. Multi-agents on the network and under-test components communicate and interact to find the performance result and the bottle neck of the system. Architecture includes the agents’ arrangement, different agent’s distribution and their running process. It is a black-box test method and could simulate several different parallel relative transactions at the same time. Test is started from link end that no configuration should be done. This architecture is proved by a real information system test.

The aim of this paper is to investigate a control method for robotic agent, operating in ubiquitous network. It is difficult for people to fulfill their tasks under dynamically changing environment. Therefore, robotic agent performs tasks in stead of people. The ubiquitous network makes it possible to connect between robotic agent and people. Because robotic agent should fulfill tasks according to the received commands from people, the adaptive control method for the agent is needed in order to do the given tasks properly. This paper introduces an adaptive tracking control method for robotic agent based on the radial based functions network (RBFN). When some commands are received through networks, the proposed method can make robotic agent possible to perform tasks under dynamically changing environment. Experimental results show that the proposed control method based on RBFN is adaptable to the environment changes and is more robust than the conventional PID control method and the neuro-control method based on the multilayer perceptron.

This paper presents a framework for controlling intelligent robots connected by communication networks. This framework provides novel methods to control coordinated systems using higher-order mobile agents. Higher-order mobile agents are hierarchically structured agents that can contain other mobile agents. The combination of the higher-order mobile agent and mobile multi-robots open a new horizon of efficient use of mobile robot resources. Instead of physical movement of multi-robots, mobile software agents can migrate from one robot to another so that they can find the most suitably equipped and/or the most suitably located robots to perform their task. Thus the framework presented here provides efficient use of robot resources. In this paper, we focus on the energy saving. We have demonstrated the efficiency by numerical experiments.

We present a plan-based control architecture for intelligent robotic agents. Many modern robot architectures adopt a hybrid approach using control-theoretic mechanisms as well as plan-based mechanisms, because of their complementary strengths and weaknesses. Our three-layered control architecture implements dependability through plan-based management of a set of distributed independent behavioral components. With navigational tasks in a complex maze, we investigate robustness of our control architecture.

The subject matter of the study are the automated guided vehicle (AGV) operation synchronisation mechanisms in flexible manufacturing systems. Since each AGV can be treated as an autonomous object capable to undertake decisions regarding its routing, entering path sectors, etc., hence a class of transportation systems considered can be seen as a class of multi-agent ones. In many practical cases transport operations are repetitive. The processes examination has to guarantee the collision-free and deadlock free AGVs flow. In this paper the problem of determination of the rules coordinating access of the vehicles to the shared travel route intervals, ensuring the collision-free and deadlock-free execution of the repetitive processes was reduced to determination of the sufficient conditions of the form of a pair (initial state, a set of priority rules). In particular the problem of searching for a pair is defined in the form of the constraint satisfaction problem (CSP) and is solved with use of the logic programming techniques.

Multi-agent technologies are essential in realizing the upcoming ubiquitous environment. In the multi-agent environment, each agent has its own set of services and stores these services in the service repository of the multi-agent system. By using this repository, the user can retrieve the most appropriate service. In this paper, we propose an efficient service repository architecture that can improve the existing repository by using context inferencing, context filtering, user-predefined policies, and the categorized tree. An advantage of the proposed architecture is that the user can obtain more suitable services than existing approaches about service discovery.

The Trading Agent Competition (TAC) simulates a challenging game environment where competing agents engage in complex decision-making activities with purpose of maximizing their profits. One of the TAC competitive scenarios is Supply Chain Management (SCM) where six trading agents compete by buying components, assembling PCs from these components and selling the assembled PCs to customers. In this paper, we briefly describe the TAC SCM environment and present the main features of the CrocodileAgent, our entry in the 2005 TAC SCM Competition. The agent’s performances in the 2005 TAC SCM, as well as in a series of controlled experiments, are discussed.

The spread of internet is so rapid that emerging e-Business is making dramatic changes in this digital economy. An important component of e-Business research is the issue of the effective marketing in the keen competition. Many agents are developed to help buyers handle the dynamic purchasing environment and reduce the complexity of purchasing data. However, while most existing agents are focused on buyers’ needs, prior research on agents lacks the sellers’ perspectives. This study is intended to demonstrate the agent system that supports analysis of buyers’ purchasing data and predicts their behavior so that sellers can develop more systematic marketing strategies.

Agent technology is emerging as a new software paradigm in the areas of distributed computing. The use of multiple agents is a common technique in agent-based systems. In distributed agent systems, it is often required for two agents to communicate securely over a public network. Authentication and key exchange are fundamental for establishing secure communication channels over public insecure networks. Password-based protocols for authenticated key exchange are designed to work even when user authentication is done via the use of passwords drawn from a small known set of values. There have been many protocols proposed over the years for password authenticated key exchange in the three-party scenario, in which two agents attempt to establish a secret key interacting with one same authentication server. However, little has been done for password authenticated key exchange in the more general and realistic four-party setting, where two clients (or, two agents) trying to establish a secret key are registered with different authentication servers. In this paper, we propose a new protocol designed carefully for four-party password authenticated key exchange that requires each agent only to remember a password shared with its authentication server.

The development of ubiquitous systems for maintenance and control of treatment systems to assist individuals in managing their medical treatment plan would provide an improved system for home healthcare. The assistance of agent-based systems to help doctors and further automate medical treatment systems is limited by medical knowledge data mining accuracy and the complexity of each patient’s health history and individual. Automation of knowledge base development for each individual patient would allow efficient personalization of each patient’s treatment plan and allow integration of the doctor’s individual diagnosis and treatment plan. This paper presents an overview of agent based technologies and describes both historical and state-of-the-art applications of agent technologies in the medical field. Current research and development activity is identified and an algorithm to address the knowledge acquisition bottleneck for diagnostic medical knowledge is present. The algorithm can reduce time consuming knowledge acquisition and allow efficient development of individually tailored medical treatment knowledge bases.