论文部分内容阅读
针对基于多Agent的航天自主运行系统,梳理了模型描述的内容和要求。为了满足多Agent之间协同规划以及Agent内部自主规划的要求,提出了活动和方案两级的规划系统描述模型。在活动一级建立与时间、资源的约束关系,实现Agent之间的规划分解、交互协商。以方案描述系统组成部分的运行状态,并通过方案衔接约束和能量参数的定义,使Agent具有局部优化和适应变化的能力。使用选择性约束和优先级等参数,使规划活动更为灵活和鲁棒。Agent知识描述语言使用类似C++的语法,使用预定义类和函数来实现规划知识的描述,方便实现过程性运算,易于扩展。通过规划模型描述的实例验证了描述语言的描述能力。模型语言的处理程序已经应用于仿真系统,可以针对应用场景进行仿真。
Aiming at the autonomous system of aerospace based on multi-agent, the contents and requirements of model description are summarized. In order to meet the requirements of coordinated planning among multi-agents and autonomous planning within the agents, a planning system description model of activities and schemes is proposed. At the activity level, it establishes the constraint relation with time and resources to realize the planning decomposition and mutual negotiation between Agents. The program states the operating status of the system components, and through the definition of the protocol convergence constraints and energy parameters, the Agent has the ability of local optimization and adaptation. Use parameters such as selective constraints and priorities to make planning activities more flexible and robust. Agent knowledge description language uses syntax similar to C + +, using predefined classes and functions to realize the description of planning knowledge, to facilitate the process of computing, easy to extend. The descriptive ability of the description language is verified through the example described by the planning model. Model language processing has been applied to the simulation system, you can simulate the application scenarios.