基于智能的行为选择机制(BSM)存在两个缺陷:不确定性环境下,不能实现自主行为选择,在相同环境条件下,缺乏行为反应的多样性.受情绪机制应用于情感计算与强化学习研究的启发,将其应用于行为选择机制研究,以解决上述问题.从而,提出三类情绪行为选择机制:情绪影响机制、情绪驱动机制、情绪涌现机制.但这些机制存在:未能将其所需涉及的基本因素进行涵盖,有系统性缺陷;各种选择机制体系结构和情绪计算模型彼此各异,没有继承性;基于设计思想,缺乏相关学科关于情绪行为理论的支撑.另外,研究方法上,也存在各行为机制之间、智能机制和情绪机制之间缺乏系统性分析和比较的问题.提出,要揭示情绪行为选择机制,应从认真挖掘动物行为学、心理学、经济学等领域有关情绪研究结果之间的内在一致性入手.同时,比较方法的使用,将是加速这一过程的有效手段.
There are two defects on intelligence-based Behavior selection mechanism(BSM): under uncertainty conditions,behavior selection can′t be achieved autonomously;under same conditions,behavioral response is absence of diversity.Being inspirited by research achievements in emotion mechanism application for emotion calculation and reinforcement learning,emotion mechanism is applied to research on BSM to settle the above problems.Therefore,the three kinds of emotional BSM(EBSM) are suggested:emotion-affected mechanism,emotion-driven mechanism,emotion-emerged mechanism.However,these mechanisms have three problems: a systematic drawback exists that the necessary elements,which BSM should have had,have not involved in;The EBSM architectures and their emotion models are different from each other,there is no inheritane;They are design-based,there are not the corresponding academic supports for them on emotion behavior mechanism.In addition,there exists a problem on the research methods that systematic analysis and comparisons have not conducted among the EBSMs,between intelligent BSMs and emotional BSMs.The suggestions are offered that revealing emotional behavior selection mechanism should be start off with carefully mining the internal consistency among relevant research results of emotion in ethnology,psychology,microeconomics,etc;meanwhile,comparative research method will be an effective means to accelerate the course.