MOOC人工智能原理学习笔记2

人工智能原理学习笔记2
一、任务环境
1.PEAS is a task environment specification, stangs for:
Performance, Environment, Actuators, Sensors.

2.Different Environment Types
Fullly observable vs. partially observable 完全可观测与部分可观测
Single agent vs. multi-agent 单智能体与多智能体
Deterministic vs. stochastic 确定性与随机性
Episodic vs. sequential 阵发性与连续性
Dynamic vs. static 动态与静态 (Semi-dynamic: 半动态 If the environment itself does not change with the passage of time but the agent’s performance score does. )
Discrete vs. continuous 离散与连续
Known vs. unknown 已知与未知

二、智能体结构
1.Agent Structure 结构

2.Agent Function 函数
决策制定的原则：
calculation of utility of individual options, deduction over logic rules, fuzzy logic, lookup table, etc.

3.Agent Programs 程序

4.Three ways to represent states for an agent

a)Atomic representation

b)Factored representation

c)Structured representation

三、智能体的主要类别
1.Simple reflex agents 简单反射智能体

2.Model-based reflex agents 基于模型的反射智能体

3.Goal-based agents 基于目标的智能体

4.Utility-based agents 基于效用的智能体

5.Learning agents 学习智能体

四、智能体视角

五、智能体的分类法