China
Africa
Explore chapters and articles related to this topic
Negotiation and Conflict Resolution
Published in Weiming Shen, Douglas H. Norrie, Jean-Paul A. Barthès, Multi-Agent Systems for Concurrent Intelligent Design and Manufacturing, 2019
Weiming Shen, Douglas H. Norrie, Jean-Paul A. Barthes
Contracting was first proposed in the Contract-Net (Smith, 1980) and then demonstrated on a distributed sensing system (see Section 3.3.2 in Chapter 3). To summarize: each agent (manager) having some work to subcontract broadcasts an offer and waits for other agents (contractors) to send bids. After some delay, the best offers are retained and contracts are allocated to one or more contractors who process their subtasks. The contract-net protocol provides for coordination in task allocation, with dynamic allocation and natural load balancing. The approach is quite simple and can be efficient. However, when the number of nodes is large, the number of messages on the network increases, which can lead to a situation where agents spend more time processing messages than doing the actual work, or worse, the system stops through being flooded by messages. Thus, various improvements to the basic contract-net approach have been proposed, such as: - sending offers to a limited number of nodes, instead of broadcasting them;- anticipating offers, i.e., contractors send bids in advance;- varying the time when commitment is decided;- allowing de-commitment (breaking commitments);- allowing several agents to answer as a group (coalition formation);- introducing priorities for solving tasks.
Agent-Driven Wireless Sensors Cooperation for Limited Resources Allocation
Published in Fei Hu, Qi Hao, Intelligent Sensor Networks, 2012
Sameh Abdel-Naby, Conor Muldoon, Olga Zlydareva, Gregory O’Hare
The Contract-Net protocol presented in [4] is a high-level negotiation protocol for communicating service requests among distributed agents. R. G. Smith considers the high-level negotiation protocols as methods that lead system designers to decide “what agents should say to each other.” And low-level protocols make system designers decide “how agents should talk to each other.” The Contract-Net protocol assumes the simultaneous operation of both agents asking to execute tasks and agents ready to handle it. The asking agents broadcast a call for proposals, and the helping agents submit their offers and then one is granted the pending task, or the session is closed. Three points are worth highlighting in the earlier approach: Linking between high-level and low-level negotiation protocols is essential when it comes to agents interacting in limited and variable resources environment. For example, when users of pocket computing devices delegate software agents to exchange and accomplish service requests on the go, the efficiency of the negotiation protocol that agents will employ is relatively increasing with the size of bandwidth a network utilizes, and the time it takes to transfer agent’s requests/messages from one location to another.A central decision-making situation may easily occur when a service seeker initiates the call for proposals, and it receives back all of the prospects offers, and the same agent is the only one who decides upon the termination of the negotiation process.In the Contract-Net, it is always assumed that two different types of agents are interacting (e.g., buyer and seller agents), which is different in the sensors scenario because they are all of the same type.
Business Improvement through Innovation in Construction Firms: The ‘Excellence’ Approach
Published in Ben Obinero Uwakweh, Issam A. Minkarah, 10th Symposium Construction Innovation and Global Competitiveness, 2002
Herbert S. Robinson, Patricia M. carrillo, Chimay J. Anumba, Ahmed M. Al-Ghassan
Negotiation is proposed as a means for agents to communicate and compromise to reach mutually beneficial agreements. The model here examines the problem of resource allocation and task distribution among autonomous agents, which can benefit from sharing a common resource or distributing a set of common tasks. Negotiation is initiated by the generation of a new task. This may occur when one problem solver decomposes a task into sub-tasks, or when it decides that it does not have the knowledge or data required to carry out the task. When this occurs, the agent issues out a call for proposal (CFP) message (see Figure 1.3). The arrival of a CFP message on the Supplier’s side causes it to move into initialization state. If the Supplier agent decides to respond, it will move into negotiation state. The negotiation protocol in this model constitutes two approaches: a simple and contractual approach. The simple approach may be viewed as a generalisation of the contract-net-protocol (CNP) (Nwana and Ndumu, 1996). In CNP, task distribution takes place through a process involving manager task announcement, followed by bids from competing contractors and finally the announcement of awards. The contractual approach involves an interactive process of offers and counter-offers in which each agent chooses a deal which maximizes its expected utility value. The bidding prices commences low and increase while the asking price start high and decrease (see Figure 1.4). The figure depicts that the supplier will offer the material at the highest desired price, and then decreases this price according to the decay function (which is specified as being linear, quadratic or cubic). The converse is true with the Buyer as shown in the figure. Thus, during negotiation, the Buyer role of the supply chain predecessor negotiates with the Supplier in the successor. It involves request message sent out to all potential suppliers, who on their part have to agree to go into negotiation or refuse to negotiate because of other commitments. The final stage of negotiation, involve the agents applying their negotiation strategies to reach an agreement. This process commences at the supply chain head, and finishes at the supply chain tail (see Figure 1.2).
A multi-agent and internet of things framework of digital twin for optimized manufacturing control
Published in International Journal of Computer Integrated Manufacturing, 2022
Qingwei Nie, Dunbing Tang, Haihua Zhu, Hongwei Sun
The most commonly used negotiation mechanisms based on multi-agent technology are the contract net protocol and game theory (Yeung 2018; Esfahani, Hariri, and Mohammed 2019). However, the original contract network protocol is communication intensive, and a heavy communication load hinders the agents to respond to unpredictable exceptions in the multi-agent system (Novas et al. 2012). The negotiation mechanism based on game theory is seldom fully implemented in the agent negotiation process, which is less mature than the mechanism based on contract network protocol (Valckenaers 2019). This is especially true for a manufacturing shopfloor in the IoT environment with the vast amount of data concurrency and exchange (Morariu et al. 2020). The improved contract net protocol (ICNP) can solve the problem of large network traffic. Thus, it is necessary to apply the ICNP to reduce the communication burden among the agents and improve the operation efficiency of the physics shopfloor.
Efficiency of task allocation based on contract net protocol with audience restriction in a manufacturing control application
Published in International Journal of Computer Integrated Manufacturing, 2018
The contract net protocol (CNP) (Smith 1980) embodies a negotiation metaphor for distributed problem solvers, or agents, to cooperate through exchange of messages (Davis and Smith 1983). Multi-agent systems (MAS) based on CNP have been applied in different areas of manufacturing including shop floor control (Parunak 1987; Shaw 1987; Cantamessa 1997; Krothapalli and Deshmukh 1999; Ünver and Anlagan 2002; Merdan et al. 2011; Komma, Jain, and Mehta 2011; Renna 2011; Borangiu et al. 2014; Renna 2015), process planning and scheduling (Macchiaroli and Riemma 2002; Jégou et al. 2006; Wong et al. 2006; Shukla, Tiwari, and Son 2008; Zattar et al. 2010; Li, Zheng, and Yang 2010; Nejad et al. 2010; Archimede et al. 2014; Barenji et al. 2017) and automated guided vehicles routing (Srivastava et al. 2008; Carpanzano et al. 2016). More recently, CNP has featured in other manufacturing applications such as workflow management (Wang and Zeng 2010), virtual/extended enterprises (Fung et al. 2008; L´opez-Ortega and de La Cruz 2010), project scheduling (Hsieh and Lin 2015), cloud computing (Sim 2012) and web service composition (Lee et al. 2012).
Design of scalable agent-based reconfigurable manufacturing systems with Petri nets
Published in International Journal of Computer Integrated Manufacturing, 2018
In Section 2, the CSP is to find a configuration with minimal number of resources to meet product demand by due date . As the models of operation agents and resource agents are described by Petri nets, the number of resources in a configuration can be represented by a marking in Petri nets. Therefore, the CSP can be formulated based on the relevant Petri nets of individual operation agents and resource agents discovered in the problem-solving process. The problem-solving process relies on collaboration and interactions of operation agents, resource agents, production control agents and scheduling agents. Interactions between different types of agents are based on the contract net protocol (CNP) (Smith 1980), which consists of task announcement, bidding mechanism, contract awarding and negotiation. The CNP used in this paper is a multi-level CNP applied between an operation agent and its upstream operation agents and between an operation agent and its relevant resource agents to optimise the solution. An operation agent announces a task to relevant agents, waits for the bids submitted by its upstream operation agents, chooses among the bids received and decides whether to award the contract to any of the bidder by solving the operation agent scheduling problem (OASP), which will be formulated later based on the structure of Petri nets.