Linear Programming Learning Notes (1) Introduction

Linear Programming Learning Notes (1) Introduction

All the resources come from Linear Programming: Foundations and Extensions by Professor Robert J. Vanderbei.
Explore the link below for further information:
LP Book Resources

Part 1

Basic Theory: The Simplex Method and Duality

Chapter 1 Introduction

Examples

• Resource Allocation
• Blending Problem(Diet Problem)
• Shape Optimization(Telescope Design) (Not quite understood)
• FIR Filter Design(Not quite understood either)

• Portfolio Optimization
  -A Markowitz Type Problem
  If we need to optimize two objectives, we could make a compromise to set a bound of one and maximize(minimize) another.
  -Efficient Frontier
  Varying risk bound $\mu$ produces the so-called ecient frontier.
  Portfolios on the ecient frontier are reasonable.
  Portfolios not on the ecient frontier can be strictly improved.

  Definitions

• Optimist and Pessimist
  -Production manager as Optimist, to maximize the product quantity.
  -Comptroller as Pessimist, to minimize the cost.
• The Linear Programming Problem
  -Decision Variables:
  $x_j, j=1,2,3,...,n$
  -Objective Function:
  $\zeta = c_1x_1 + c_2x_2 + … +c_nx_n$
  In real-world problems are most naturally formulated as minimizations(since real-world planners always seem to be pessimists), but when discussint mathematics it is usually nicer to work with maximization problems.
  -Constraints:
  $a_1x_1 + a_2x_2 + ... + a_nx_n = \left\{\begin{aligned}\leq \\= \\\geq\end{aligned}\right\}b.$
  -Mathematically preferred presentation:
  maximize $C^TX$
  subject to $AX \leq B, X\geq0$
  where $C \in {\rm I\!R{^n}}, X \in {\rm I\!R{^n}}, A\in {\rm I\!R} ^{m\times n}, B \in {\rm I\!R} {^m}$
  m constraints, n variables
  -Infeasible problems
  -Unbounded: Feasible solutions with arbitrarily large objective values.