What‘s the utility theory in artificial intelligence?

题意 ：人工智能中的效用理论是什么？

问题背景：

It has the book Artificial Intelligence: A Modern Approach by Stuart Rusell. I am currently engaged in chapter 16, which is titled "Making simple decisions", and I'm struggling to grasp the core concept of utility theory. Could you elaborate with a specific example to help me better understand this concept?

我的手中有一本《人工智能：现代方法》，作者是Stuart Rusell。我在研读第16章‘做简单决策’部分时遇到了困难——我对效用理论的核心概念尚未完全 grasp. 请问能否提供一个具体的案例来说明这一概念？

问题解决：

从本质上说，效用理论的核心概念非常容易理解：agent 对可能结果的偏好可用函数来表示；将这些结果映射到实数域；当数值越高时，该agent对相应结果越喜爱。其中这种函数被称为效用函数。

效用理论的核心理念极为简洁：其主要观点是表明个体在面对不同可能的结果时表现出的偏好程度；这种偏好可以通过一种将这些结果转化为实数值的机制得以量化表示；具体而言，在这种机制中数值越高就意味着个体越倾向于选择对应的选项。这种转换关系被称作效用函数。

For example, we could say that my utility for owning various items is:

例如，我们可以说我对拥有各种物品的效用为：

 u(apple) = 10

    
 u(orange) = 12
    
 u(basketball) = 4
    
 u(macbookpro) = 45

Economists (typically) perceive humans as utility-maximizing agents, which can be understood as our constant pursuit of maximizing the internal utility function.

经济学家（一般情况下）把人类视为效用最大化的目标。换句话说，在这种观点下，我们总是致力于追求和实现自己的效用函数的最大化。

Once you obtain these numerical values, they can be combined with probabilities to discuss the expected utilities of optimal strategies, discounted future benefits, and various additional interesting aspects.

当拥有这些数值时

To learn more, obtain a copy of a textbook about game theory, instead of reading the first chapter.

如果你有兴趣获取更多信息, 可以学习博弈论基础知识, 或者参考该领域的权威教材.this agents book.(http://multiagent.com/2010/02/multiagent-systems-textbook.html "this agents book.")