【作者】 何娟；

【导师】 谢天保；

【作者基本信息】 西安理工大学 ， 管理科学与工程， 2007， 硕士

【摘要】 日益加剧的市场竞争和个性化的客户需求使得现代制造企业正面临着前所未有的挑战。企业经营的市场环境发生了巨大变化,企业之间由单纯产品质量、性能方面的竞争转向企业所在的供应链之间的竞争,影响企业生存、发展的共性问题是目前竞争环境、顾客需求因素变化太快,而企业自我调整适应的速度跟不上,通过敏捷制造来达到敏捷竞争,将是企业参与国际竞争的主要形式。本文正是在这种背景之下,结合智能计算领域的Agent概念,建立了把Agent技术与合同网投标机制结合构建了敏捷制造生产任务调度模型;在信息不完备模型下,提出动态调度优化算法,结合遗传算法求取敏捷调度的最优方案。论文对Agent技术,多Agent系统,敏捷制造调度系统结构、优化调度算法、程序实现等技术进行了深入对比分析研究。针对敏捷制造单元调度问题的动态性、开放性和信息不完备性等特点,论文构建了基于多代理生产组织和运行模式的生产任务调度系统框架结构,研究了敏捷制造环境下的不完备信息模型,提出基于CNCP扩充网协议的自组织招标机制以及基于遗传算法的调度优化模型。最后,论文作出仿真模拟试验并在JADE平台下实现系统。实验证明:在不完备信息模型下采用该算法可求取调度目标的最优解。敏捷制造的特点是能够快速、及时的响应市场变化,这就要求企业的生产系统能够快速重构来自不同企业的制造单元,随着任务的生成而创建,随着任务的完成而解体。本文构建的多代理生产组织和运行模式的生产任务调度系统框架结构,可以很好的控制、协调各代理间的活动,有效配置资源;而基于CNCP协议和自组织招标机制下的敏捷制造调度系统可以解决敏捷制造中的动态性和开放性的问题。在实际的合作伙伴的选择过程中,系统结合遗传算法对选择方案进行优化,得到最优可行解,最终解决敏捷制造单元调度的信息不完备性、动态性和调度结果优化等问题,为敏捷制造合作伙伴的选择提供了良好的支持系统。更多还原

【Abstract】 The stringent market competition and individualization requirements make today’s manufacturing enterprises facing drastic challenges. Enterprises’market environments have hugely changed. Enterprises have competed from competing product’s quality、capability to supply chain competition;Changeable competition environment and client requirement influence enterprise’subsistence and development, but enterprise speed-autoadjusting is slow.through agile manufacturing to agile competition, which is primary form of enterprises taking part in international competition. Just under this background, this dissertation constructs agile manufacturing scheduleing model based on Agent technique and Contract Net’s bidding mechanism with agent concept in computer field; under imperfect information model, dynaically optimized scheduling algorithm is put forward and optimized agile scheduling’s project is obtained with the genetic algorithm.There is an overall and deep study on the application of Agent technique、multi-Agent system、agile manufacturing scheduling system framework、optimized scheduling algorithm、program implementing technique. In terms of the characteristics of units of agile manufacture such as dynamics, openness and imperfect information, the paper builds the structure for the dynamic scheduling system of production tasks founded on production organizations and operating modes of multi-agent, and explores the imperfect model under the circumstance of agile manufacture. In addition, it puts forward the self-organized bid mechanism based on the CNCP-extending network agreement and the optimized model for dynamic scheduling based on the genetic algorithm. Finally, it makes emulational simulative experiment and implements the system under JADE flat. Experimental results have evidenced that the algorithm can obtain the most optimal solution for the scheduling target in the case of imperfect information model.Agile manufacture’s characteristic fleetly and duly response to market change, which requires that corporation’s production system can fleetly recombine manumacture unit in different corporation, establishing with task’s building and disjointing with task’s fulfilling. This paper builds the structure for the dynamic scheduling system of production tasks founded on production organizations and operating modes of multi-agent, which can commendablycontrol and cooperate Agent’s activities as well as effectively collocate resources; agile manufacture scheduling system based on CNCP and Self-organized bidding mechanism can solve dynamics, openness in agile manufacture. In the parter’s choice course, the system optimizes project with genetic Algorithm and obtains the optimized solution, which finally solves agile manufacturing unit’s dynamics, openness and imperfect information and provides favorable sustaining system to agile manufacturing’s parter.更多还原