Contract Net Protocol

About: Contract Net Protocol is a research topic. Over the lifetime, 606 publications have been published within this topic receiving 13729 citations.

Implementing contract net in tuple space models.

Abstract: Coordination is essential in distributed computing, which involves a group of loosely coupled agents cooperating in order to solve problems in a decentralised fashion. One of the popular approaches of coordination is the generative tuple space based communication, where agents interact via a shared data space. This paper discusses the implementation of Contract Net protocol (CNP) in a number of shared data space models, beginning with LINDA, and followed by several of its more secure variants. And finally, taking into account the importance of secure conversation in the process of task contracting, we propose the combination of Scope (a variant of LINDA) and access control attributes as the most appropriate model to be used as the mechanism in implementing CNP in the shared data space communication model. Apart from its tagging ability to limit visibilities and streaming property, the other significant advantages offered by Scope is its dynamicity (due to the scope-based operations), which is particularly useful in contracting several related tasks, where agents’ visibilities can be tailored (widened or limited) according to need. Incorporating attributes in the model provides a more secure contracting, without losing the flexibility offered by Scope.

Research on contract net schedule algorithm based on medium pattern

Abstract: Contract net schedule algorithm based on medium pattern was proposed. In this algorithm, manager provides information service for contractor and task agent as medium during releasing contract. Contractor and task agent evaluate each other according to its evaluation function and then select the suitable object. Compared with traditional contract net, the algorithm in which task is regarded as agent reduces the communication and load between manager and contractor. Besides, task executing time is cut down. At last, taking the production schedule for example verifies this algorithm. Tests showed that amount of communication in manager is less and the task is assigned more effectively by the contract net schedule algorithm based on medium pattern.

Design of contract net protocol-based distributed image processing system

Abstract: In the current era of large scale distributed computing, users voluntarily donate computation power of their idle computers or desktop machines. However, there is an urgent need of a utility-based distributed problem solving environment to attract more users. Related to applying the utility concept, we are working on design and implementation of a distributed image processing system that uses: 1 contract net protocol for task assignment; 2 concept of utility-based distributed computing; 3 task graph concept for application design. In this paper, we present our results of the design aspects of the system. We propose the design of contract net protocol-based distributed image processing system and analyse characteristics of the proposed system. We evaluate and select available options for implementation of the system. The proposed system allows transparent execution of task graph-based applications in a distributed environment.

Negotiation Protocol Comparison for Task Allocation in Highly Dynamic Environments

Abstract: A robust negotiation protocol is required for a multi-agent simulation involving two adversarial teams in a highly dynamic and hostile environment. In this environment agent failure is possible due to a number of circumstances such as running out of fuel or being destroyed by other agents. This paper compares three existing negotiation protocols: the Contract Net Protocol (Smith, 1980), the Distributed Contract Net Protocol (Cano and Carbo, 2006), the Extended Contract Net Protocol (Aknine et al., 2004) and two protocols developed by the authors (termed herein theSimple' andHybrid' protocols). The objective of this paper is to determine which protocol is best suited to our application in terms of scalability, robustness against agent failure, communication overhead, and response time. To evaluate these negotiation protocols an experiment was conducted, involving three different test cases, which varied the availability of agents at different stages of the negotiation process. In these test cases a team of software agents (theblue team') were tasked with destroying a number of stationary targets (thered team'). The experimental results showed that the Contract Net Protocol (CNP) was suitable for low risk environments due to its low communication overhead, while the Distributed Contract Net Protocol (DCNP) was more suitable for high-risk environments due to its greater robustness against agent failure. However, this robustness was achieved at the expense of greatly increased communication. An alternate approach that showed promising results was to use a Hybrid protocol that switched between CNP and DCNP depending on the environment. Additional work is required to develop the Hybrid protocol further.