Based on the analysis of shortcoming and limitation in traditional contract net protocol , the author provides a new cooperation strategy - improved contract net protocol , which improves the efficiency of the system and the quality of the task completed by introducing evaluated parameters and expanding the statements of tenderee and bidder . also , a bidder - selected strategy is presented . selecting a proper interaction way and expanding kqml can make kqml support high - level , complicated and multi - bout interaction and cooperation , and can realize the improved contract net protocol 为实现改进的合同网协议，还需要对agent的通信方式进行选择并需要对agent的交互语言? kqml进行原语级的扩充，以弥补kqml中有关任务交互原语的缺乏，使kqml支持agent间关于目标、任务的高层次、复杂、多回合的交互协作，进而实现改进的合同网协议。
3 . aimed at the distributed decision - making process of production planning based on negotiation , the problems about priority of orders , order / task decomposition , order / task allocation etc are developed . moreover , aimed at conflicts which occur in the process , an improved contract net protocol and some negotiation - based strategies are presented 3 .针对gr - hppcs框架下基于协商机制的生产计划分布式决策过程,分别研究了订单优先权、订单/任务分解、订单/任务分配等决策问题,并且针对决策过程中可能出现的冲突提出了基于改进合同网协议的分级协商模型与协商策略,以解决系统运作过程中需要有效解决的问题。
We analyze agent cooperation methods now commonly in use , and , on the basis of ecology and experiences , we propose a competitive price auction contract network protocol . while using the same network traffic volume as the simple auction , our strategy of second overbid competitive price auction can achieve the same effect as the traditional auction . further , we improve traditional contract network protocol , introduce constraints into message transmit , and use fell back punishment and credit when an agent cannot complete the intended task , consequently the traffic volume and the complexity of systems are considerably reduced 本文分析了现在通常采用的agent协同方法，提出基于生态学和经验的竞价拍卖合同网协议，采用的选择次高价竞价拍卖策略能在与简单拍卖相同的通信量上达到类似于传统竞价拍卖的效果，对传统的合同网协议进行改进，在消息传播中加入约束机制和对agent不能完成任务时不采取转变角色重新招标而采用违约惩罚和信用机制，减少了系统的通信量和系统的复杂度，最后在此基础上实现一个基于fipa标准的多agent电子商务原型系统。
Many researchers presented corresponding models for multi - agent negotiation framework . rosenschein introduced a negotiation model using the idea of game theory , sycara constructed a model based on the labour and capital negotiation , kraus created a common negotiation framework considering sociality and cooperation in negotiation , sandholm gave a commitment based negotiation mechanism . another model given by jennings took the changing of the mental state and joint - commitment into account 许多学者提出了一些理论和方法，包括rosenschein等人的基于对策论的理性协商理论， r . smith等人提出的合同网理论， sycara的以劳资谈判为背景的协商理论， kraus的最佳平衡协商理论， sandholm的多agent核结盟理论以及jenning等人基于协商中思维状态变迁的联合承诺理论等等。
Finally , a fault diagnosis system of civil aircraft was developed on the jade platform , and realized the distributed cooperation diagnosis of many cbr - agents based on contract network , as the combination of the cbr method and the agent technology , the system has the properties of collective intelligence , good robustness and better fault tolerance 最后,在jade平台上实现了采用合同网方式的多个cbragent分布式协作诊断,开发完成了民用飞机故障诊断系统, cbr推理方法及agent技术相结合的使用,使得该系统具有集体的智能性、较好的健壮性与容错性。
From the simulative experiment and comparing of different cooperative model , we find that toac cooperative model can solve multi - agent social organization , social convention , social evolution , coalition strategies , coalition stability , collaborative problem solving and task allocation in open and distributed multi - agent systems ; can organize multi - agent cooperation , enhance system dependability , reduce resource cost and system communication bottleneck problem ; can share the task and result of multi - agent , and can reach the global optimal of multi - agent utility in task oriented domains ; from time performance , toac model is better than cnp ( contract net protocol ) of r . smith , blackboard cooperative model of newell and the result share model of smith and davis 通过仿真试验和结果分析，发现本文提出的toac模型，该模型能够有效的解决在开放、分布的mas环境中，多agent的社会组织、社会规范、社会演化、联盟策略、联盟稳定性、合作问题求解和任务分配，能够有效的组织多agent间的协同、提高系统的可靠程度、降低资源开销和系统通信瓶劲问题，能够实现多agent任务和结果共享，并能够达到面向任务领域多agent联盟效用的全局最优。时间性能上toac模型要优于r . smith的合同网协同模型、 newell的黑板协同模型，以及smith和davis的结果共享协同模型。
