Catálogo de publicaciones - libros

This study deals with the problem of supporting negotiation on prices and due dates of multiple orders between a manufacturer and its suppliers in a make-to-order supply chain. A new negotiation agenda with two phases is proposed based on the fact that relationship between the partners is both cooperative and competitive. In the cooperative phase a mediator implementing Simulated Annealing is incorporated to help the manufacturer and the supplier search tentative agreement of due dates which minimizes the total supply chain cost. Then, they adjust the reservation value and aspiration value of price accordingly based on the idea of integrated-utility. They bargain on the price issue using concession based methods in the competitive phase. The proposed negotiation agenda and approach can achieves near-optimal social welfare and reach win-win solution for negotiation agents. Result of a numerical example is reported.

Reframing is a sub-type of interest-based negotiation strategy that enhances bargaining by allowing the negotiators to ask for the underlying goal of the negotiation and propose alternative plan(s) which may entail a deal on alternative issues. This paper (i) presents a negotiation protocol that support both alternate offers monotonic bargaining and reframing and (ii) gives a fully computational and reproducible specification of bargaining and reframing capable negotiation agents using the 3APL agent language.

In a planetary rover system called “SMC rover”, the motion coordination between robots is a key problem to be solved. Multiagent reinforcement learning methods for multirobot coordination strategy learning are investigated. A reinforcement learning based coordination mechanism is proposed for the exploration system. Four-robot climbing a slope is studied in detail as an instance. The actions of the robots are divided into two layers and realized respectively, which simplified the complexity of the climbing task. A Q-Learning based multirobot coordination strategy mechanism is proposed for the climbing mission. An OpenGL 3D simulation platform is used to verify the strategy and the learning results.

The traveling salesman problem (TSP) is a very hard optimization problem in the field of operations research. It has been shown to be NP-hard, and is an often-used benchmark for new optimization techniques. This paper pro- poses an improved multi-agent approach for solving large TSP. This proposed approach mainly includes three kinds of agents with different function. The first kind of agent is conformation agent and its function is generating the new solution continuously. The second kind of agent is optimization agent and its function is optimizing the current solutions group. The third kind of agent is refining agent and its function is refining the best solution from the beginning of the trial. At same time, there are many sub-agents in each kind of agent. These sub-agents accomplish the task of its superior agent cooperatively. At the end of this paper, the experimental results have shown that the proposed hybrid approach has good performance with respect to the quality of solution and the speed of computation.

The goal of this paper is to propose a way to register agents so that makes users or autonomous agents to find target agents more conveniently and efficiently. In order to achieve this goal, we have survey the key technologies such as UDDI, ebXML, FIPA DF. As the results, first we have designed a new agent registry/repository system based on ebXML technologies. Next, we introduce our agent registry/repository system for wide area network and describe how to use our system for registering agents and searching target agents for communications. Finally, we give an example to illustrate a typical scenario in which user receives the results from our proposed agent registry/repository system.

The existing schemes based on ant agents don’t take into account the impact of the imprecision of network state information on routing performance. In this paper, we design a novel ant agent-based multicast routing algorithm with bandwidth and delay guarantees, called QMRA, which works for packet- switching networks where the state information is imprecise. In our scheme, an ant uses the probability that a link satisfies QoS requirements and the cost of a path instead of the ant’s trip time or age to determine the amount of pheromone to deposit, so that it has a simpler migration process, less control parameters and can tolerate the imprecision of state information. Extensive simulations show our algorithm can achieve low routing blocking ratio, low average packet delay and fast convergence when the network state information is imprecise.

A fundament tradeoff issue observed by Ratnasamy is a hotspot in designing distribute hash table(DHT) in P2P network. Three constant-degree systems had proposed recently, but the common weakness of them is handle node leaving without inform its neighbors in advance, and optimize the degree and load balance is another question. In this paper, a constant-degree system has proposed named Cactus which based on the 2-tree and CCC hypercube. Its number of neighbor is 6, and the time complexity of key lookup is O(d) when number of nodes is no more than d*2d, d is the degree of CCC. In this paper, we will introduce the topology, routing algorithm, node join and leave for Cactus. The experimentations show that Cactus is better in optimizing the degree and load balance, faults tolerate and no worse in other performance compare to other constant-degree system.

P2P systems have high scalability, robustness and fault tolerance because of its no centralized server and the network self-organized itself. As the standard protocol for VoIP and NGN, SIP is a general protocol for establishing and controlling multimedia session. In this paper, we proposed a peer to peer based realtime streaming media sharing system using SIP mechanism, which uses Chord as the underlying connection architecture. We add the media agents sever to support real time streaming media downloading. We have advanced peer and piece selection algorithm to solve the real time media stream transmission. The MPSS system architecture supports user registration, media resource location, media streaming session establishment, and realtime media streaming playback online. At last, we give a simulation and test result.

Context aware system provides a process of context acquisition and reasoning to the context-aware service and enables developers to implement of these applications easily. Unfortunately, previous studies have not been able to support the user mobility, nor the service mobility efficiently. Both, however, should be supported in a ubiquitous computing environment. Therefore, this study proposes a dynamic context aware service model, which supports the management of context information and service mobility among service gateware. We also have designed a middleware based on this model. As the middleware is implemented on the OSGi framework, it can cause inter-operability among devices such as computers, PDAs, home appliances and sensors. The proposed middleware has advantages of interoperation and integrating the heterogeneous devices by applying these standard interface technologies to this middleware.

In this paper, we consider an e-Commerce system that supports buying and selling goods or services over the Internet using Mobile Agent(MA) technologies. MAs are active, autonomous, and self-replicable software objects containing both computational logic and state information. When MA technologies are applied to e-Commerce environments, the MAs can search for a product on behalf of the buyer. The buyer constructs a group of MAs, and dispatches each agent in the group to the supplier sites. Each buyer agent compares price information of goods which are suggested by the suppliers. Eventually, all MAs exchange their price information cooperating to determine the most suitable item. In the selection process, if the price information which is exchanged among the MAs is modified by malicious entities, the user may not select intended goods. In this paper, we propose a group session key generation scheme called PGKA(Partial Group session Key Agreement) for secure communication among the MAs. The PGKA has no sponsors, controllers, or any other members charged with special duties. The main idea in PGKA is to distribute the computation of the group session key generation among the members, which makes the scheme suitable for fully distributed environments. In addition, the cost of the key reconstruction is very low in this scheme, because only the seed value is required to be transmitted for the reconstruction process.