This preview shows pages 1–2. Sign up to view the full content.
MIE376 Mathematical Programming
Lecture Notes
Daniel Frances © 2011
1
Stochastic Programming – LP with Resourse
The next frontier in LP is how to incorporate uncertainty into LP. After all it is naive to
believe that in the standard LP,
Max cx s.t. Ax
≤
b x
≥0, the values of c, A and b are known
with certainty.
The greatest challenge is how to deal with uncertainty in the constraints – it is easy to show
that as long as we are risk neutral it is conceptually acceptable to replace the c’s with their
expected values. But how do we deal with the constraints data.
Two approaches are possible
a. Chance Constrained Programming – an approach that has fallen out of grace
b. LP with Recourse – very popular in finance and energy – now known simply as Stochastic
Programming.
What is assumed here is that the problem includes at least two stages:
1.
A decision making stage, during which there is uncertainty.
2.
A recourse stage, during which any resulting constraint violations can be fixed.
In these notes we will only deal with two stages, although the same principles can be easily
expanded to formulate problems with more stages. Unfortunately the number of variables
grows exponentially with the number of stages, and quickly becomes numerically intractable.
We will solve the recourse programming for discrete distributions both without and with
stochastic programming, in the context of an agricultural setting.
You own 500 acres of land in which to plant wheat (w), corn (c) and beats (b) with planting
costs are 150 $/acre for wheat, 230 $/acre for corn and 260 $/acre for beets
To feed your cattle you require 200 T of wheat and 240 T of corn.
The yield of each crop is uncertain with expected values of 2.5, 3 and 20 T/acre for what, corn
and beets respectively. Purchasing cost of wheat is 238 $/T and 210 $/T for corn. Any excess
wheat can be sold for 170 $/T, corn for 150 $/ton, beets for 36 $/ton (up to 6000 T), beets for
10 $/T (beyond 6000 T).
Clearly this problem has the two required stages.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview. Sign up
to
access the rest of the document.
This note was uploaded on 04/20/2011 for the course MIE 376 taught by Professor Daniel during the Spring '11 term at University of Toronto Toronto.
 Spring '11
 Daniel

Click to edit the document details