Topic 1: Preference and Utility1. Preference relations2. Utility functionRepresentation theorem:Convexity of Utility/Preference/Indifference curve:3. Summary
Preference, At least as good as: 
Strict preference: 
Rational: The preference relation 
Completeness: for all 
Transitivity: For all 
If  
if 
Additional assumptions on preference.
Continuity : For every converging sequence of pairs of consumption bundles 
A violation: Lexicographic preference(字典preference)
Local Non-satiation: For 
Monotone:
(MWG) The preference relation 
(JR) The preference relation 
Convexity: for 
Strict convexity: 
Indifference curves demonstrate diminishing marginal rate of substitution
preference is convex, but utility function is quasi-concave.

Binary preference relation 
Function 
Rational & continuous preference 
" 
“ 
Proof
We want to show that
such that if and only if . Let
, where , and If for
, we have exists and is unique, then is the utility function that we are looking for. Check for existence:
Define
and , and and are both closed. By completeness,
, there exist such that . Suppose
, then it implies that two preference cannot be simultaneously and , which violates completeness of preference. Therefore, we know that . Check for uniqueness:
We can denote
, and . if
, then, there must be some and , say , and such that , which is impossible!
Therefore,
must be unique. 
Example - Lexicographic preference
This preference is rational, but not continuous. Hence, we cannot express this preference with a utility function.
Let
, and 
. but
. 
The following statements are equivalent:
Utility function is quasi-concave
Preference is convex
Indifference curve is convex
The upper contour set is convex
For every rational (complete and transitive), continuous, strongly monotonic and (strictly) convex preference, we can find a continuous, strictly increasing and (strictly) quasi-concave utility function to represent it.
Utility functions are ordinal, and positive monotonic transform of a utility function can represent the same preference.