文档介绍：河北工业大学硕士学位论文
基于多 Agent 的协作运营调度方法研究
摘要
随着市场竞争的全球化,制造业面临着更为激烈的全球化市场竞争,企业之间的协作
变得越来越重要。出现了以全球化分布式协作制造为手段,强调知识经济和大批量定制生
产的协作制造模式。协作运营网络是根据协作制造哲理组织的、以满足顾客需求为导向的
价值增值网络。协作运营调度问题是协作运营网络的核心环节,然而由于现实中各协作实
体具有相对独立的经济利益和复杂的协作与竞争关系,这个问题十分复杂。
本文构建了基于多Agent的协作运营调度系统,根据调度过程,将协作运营调度分为预
测调度和反应调度两个阶段。在预测调度优化阶段,构建了数学模型,并引入蚁群算法进
行求解。并通过仿真验证了调度模型的正确性,以及验证了算法的可收敛性和可调整性。
在反应调度阶段,协作实体通过一个两阶段的决策过程解决延迟事件,分别为模糊决策和
补偿协商过程,并用一个例子说明决策过程。
关键词: 协作运营,调度,多 Agent,蚁群算法
河北工业大学硕士学位论文
RESEARCH ON MULTI-AGENT BASED COLLABORATIVE
OPERATION SCHEDULING METHOD
ABSTRACT
With the development of the globalization of petition, the manufacturing
industry faced with a drastic global petition. The collaboration among the
companies is ing increasingly important. Collaborative manufacturing is a new
manufacturing model, and the distance of collaborative entity is cross-anization,
trans-city or transnational. The coollaborative work is the value work
which sets up according to the collaborative manufacturing philosophy. Coollaborative
operation scheduling is always the key part in coollaborative work. Because each
coollaborative entity has the relatively independent economic interest and plex
cooperation relations, this question is plex.
This thesis has constructed the multi-agent based coollaborative operation scheduling
system. According to scheduling process, the coollaborative operation scheduling divides into
two stages: predictive scheduling and reactive the phase of predictive scheduling,
establish the mathematical model, ant colony algorithm is applied. Exactness of optimization
model and characters of convergence and rectification of the algorithm ultimately gets validated
through a simulation. In the phase of reactive scheduling, the collaborative entity solves the
problem through a two-stage decision-making process: the fuzzy decision-making process and
pensatory negotiation take a case to explain