模拟人脑工作机制的机械产品概念设计行为原理启发求解方法  

Brain-like Behavior Effect-inspiring and Principle Structure-combining Method for Mechanical Product Conceptual Design

作　　者：娄山河 冯毅雄[1] 郑浩 胡炳涛 洪兆溪 谭建荣[1] LOU Shanhe;FENG Yixiong;ZHENG Hao;HU Bingtao;HONG Zhaoxi;TAN Jianrong(State Key Laboratory of Fluid Power and Mechatronic Systems,Zhejiang University,Hangzhou 310027;Hangzhou Innovation Institute,Beihang University,Hangzhou 310051)

机构地区：[1]浙江大学流体动力与机电系统国家重点实验室,杭州310027 [2]北京航空航天大学杭州创新研究院,杭州310051

出　　处：《机械工程学报》2023年第24期344-358,共15页Journal of Mechanical Engineering

基　　金：国家自然科学基金(52105281,52130501);浙江省重点研发计划(2023C01214)资助项目。

摘　　要：机械产品概念设计是现代设计理论研究的热点问题,其核心是功构映射,即根据既定的功能需求,以行为效应作为中间载体,求解满足设计约束的原理构件组合。以概念设计的功构映射理论作为基础,利用设计师大脑认知过程的神经机制作为依据,提出模拟人脑工作机制的机械产品概念设计行为原理启发求解方法。分析前额叶-海马体神经回路中设计经验知识的记忆启发机制,并以此为基础构建行为效应重构的记忆启发强化学习模型,利用多巴胺调控的Q学习在状态转移概率和设计环境回报函数未知的情况下实现行为效应的重构认知传递。同时,分析小脑模块化神经结构的反射调控作用,构建以全连接多层感知机和循环神经网络为模块的约束满足神经网络,形成复杂设计约束下的操作性条件反射判断设计空间中的约束可解性,为原理构件的组合求解提供设计依据。以电梯曳引系统的概念设计为例验证所提出方法的有效性和可行性。The research of mechanical product conceptual design is a hot topic in modern design theory.Its core is function-structure mapping,which aims to solve the combination of principle structures satisfying the design constraints according to the functional requirements and using behavior effect as the intermediate carrier.Based on the function-structure mapping theory and the neural mechanism of the designer’s brain cognitive process,a brain-like behavior effect-inspiring and principle structure-combining method is proposed.The memory inspiration mechanism of the frontal lobe-hippocampus neural circuit is revealed.Then a memory-inspired reinforcement learning model for behavior effect refactoring is constructed.The dopamine-regulated Q learning is utilized to realize the process of behavioral effects restructuring and cognitive transmission.Moreover,the reflex regulation mechanism of the modularized cerebellum is revealed,and a constraint satisfaction neural network is constructed based on fully connected multilayer perceptron and recurrent neural networks.This can form the operant conditioning under complex design constraints to judge the constraint solvability,and then provide combinatorial solutions for principle structures.A case study is conducted to illustrate the feasibility of the proposed method.

关 键 词：概念设计 功构映射 设计认知 强化学习 约束满足神经网络 

分 类 号：TG156[金属学及工艺—热处理]