Stochastic Optimization ESI 6912 - PowerPoint PPT Presentation

1 / 16
About This Presentation
Title:

Stochastic Optimization ESI 6912

Description:

ESI 6912, URYASEV, NOTES 3. 9. Objective and First Period Constraint ... ESI 6912, URYASEV, NOTES 3. 14. Value of Stochastic Solution. Expected value (EV) solution ... – PowerPoint PPT presentation

Number of Views:39
Avg rating:3.0/5.0
Slides: 17
Provided by: michaelza3
Category:

less

Transcript and Presenter's Notes

Title: Stochastic Optimization ESI 6912


1
Stochastic OptimizationESI 6912
NOTES 3 FINANCIAL PLANNING
  • Instructor Prof. S. Uryasev

2
Financial Planning
Financial Planning
3
Notations
  • to invest
  • years before college
  • tuition
  • number of investment
    periods
  • investment index
  • ( , - stocks
    , - bonds)
  • period number

4
Notations (contd)
Utility
Wealth
Utility function of wealth at year for a
goal
5
Notations (contd)
  • return on
    investment
  • decision
    vector
  • space of random elements
  • scenario number
  • probability of
    scenario

6
Scenario Tree
Tree of scenarios for three periods
1
Scenario Scenario
2
3
4
5
6
7
8
7
Scenario Representation
Returns (scenario representation)
for
otherwise
8
Tree Representation of Variables
Decisions in nodes
decisions decisions decisions
Similar, we denote returns
9
Objective and First Period Constraint
First-period constraint budget constraint
invest initial wealth
10
Constraints for Further Periods
Constraints for periods
, for each
Constraints for period
11
Model Data
  • initial wealth
  • years before college
  • surplus reward
  • storage penalty

12
Stochastic Programming ProblemTree
Representation
s. t.
13
Optimal Solution with Three-Period Stochastic
Problem
14
Value of Stochastic Solution
Expected value (EV) solution
Average return on stocks 1.155 each period
Average return on bonds 1.13 each period
Optimal investment strategy buy only stocks
EV -3.84
Stochastic solution (recourse problem (RP))
RP -1.52
VSS
VSS RP - EV -1.52 - (-3.84) 2.32
15
Stochastic Programming ProblemScenario
Representation
s. t.
set of scenarios with the same random outcomes
up to time .
16
Algorithms
- nonanticipativity constraints links
scenarios - problem can be decomposed without
these constraints
Write a Comment
User Comments (0)
About PowerShow.com