Information and Decision Services - Summary

Doubt is not a pleasant condition, but certainty is an absurd one - Voltaire
References Back

Decision Characteristics:

  • at least two courses of action available, only one can be taken
  • one or more future scenarios envisioned or assumed which, along with course of action, can influence the outcome
  • apply a decision process using the courses of action and future scenarios as input
  • the output of the decision process is the course of action to be taken.

Decision making is an integral part of the planning and control process. If the decisions to be made have been anticipated by the planning process, that is an indicator of how effective planning was. If too many decisions are required that were not anticipated during planning, the planning process may need to be refined.

Decision Process Control:

  • objective vs. bounded rationality
  • routine vs. non-routine decisions
  • qualitative vs. quantitative decision criteria
  • individual vs. group decision process
  • level of outcome certainty.

Group decision making is used often in situations with qualitative decision criteria. Group decision making has several advantages over individual decision making:

  • group interaction triggers new ideas, insights and strategies
  • groups tend to recognize incorrect solutions more quickly
  • groups have a more accurate memory
  • groups facilitate higher motivation
  • group participation helps broaden commitment for decision implementation
  • groups make less conservative decisions than individuals.

Disadvantages of group decision making include:

  • longer time to reach a decision
  • poor depth of commitment to implement decision
  • hard to fix responsibility for decision
  • group polarization can lead to extreme decisions and courses of action
  • group dynamics can overwhelm expert knowledge of an individual.

Decisions under Assumed Certainty
The outcome is assumed to be insensitive to future scenarios.
Successful decision criteria include:

  • if all alternatives have a fixed level of inputs (costs), maximize outputs (benefits)
  • if all alternatives have a fixed level of outputs, minimize inputs
  • if neither inputs nor outputs are fixed, maximize the difference.
Example:
You have a $5000 windfall to invest for retirement in 20 years. The bank offers certificates of deposit with interest compounding at 10% compounded annually or 8% compounded quarterly. Where should you invest?
Solution:
This is a fixed input problem, so maximize the output or future value of $5000.
F=P(1+i/m)^(n*M)
F(10,1)=$33,637.50 and F(8,4)=$24,377.20. Choose the first alternative.

This is a basic problem in Engineering Economic Analysis. In Engineering Economic Analysis, investments are analyzed in the context of the Time Value of Money

Assuming continuous variables, examples of successful decision processes for decisions under certainty include unconstrained optimization of the form:

Maximize(F( x1, x2,..., xn))

where the course of action is the simultaneous solution to the set of equations:
dF/dx(i)=0

For continuous variables, constrained optimization takes form:

Maximize(F( x1, x2,..., xn))

subject to:

c1( x1, x2,..., xn))<=0
 c2( x1, x2,..., xn))<=0
 .
.
.
cm( x1, x2,..., xn))<=0

the method of Lagrange multipliers is applied and the course of action is the simultaneous solution to the set of equations:
dF/dx(i)=0, dF/dLambda(j)=0
A simple form of constrained optimization problem, the linear program, is illustrated in Babcock.

Decisions under Risk
All reasonable future scenarios are known and probabilities can be assigned. Successful decision criteria include:

  • optimize the expected value or mean outcome
  • if two alternatives have the same mean outcome, minimize the variance
  • optimize a combination of the mean and variance of the outcome (example: signal-to-noise ratio)
  • optimize assuming the most likely future state
  • maximize a probability achieving an aspiration level
  • maximize the expected utility of the outcome.
The decision matrix model is convenient for independent decisions. Decision trees are used for a series of dependent decisions.

Decisions under Uncertainty
All reasonable future scenarios are known and probabilities can't reasonably be assigned. Successful decision criteria include:

  • Rationality (equally probable)
  • Pessimism (maxi-min)
  • Optimism (maxi-max)
  • Savage (mini-max regret)
  • Hurwicz (coefficient of optimism)
Dynamic Decision Making
Dynamic decision environments (many explanatory variables and risk sources; complex variable interaction; many decision makers; time-critical decisions) may benefit from computer-based tools including:
  • expert systems
  • simulation
  • groupware
  • physical modeling
  • enterprise modeling
  • integrated databases
  • query and decision support systems
  • visualization systems
Additional References:
  • Babcock, D.L. Managing Engineering and Technology, 2nd ed., Prentice Hall, Upper Saddle River, 1996.
  • Gass, S.I., Decision Making, Models and Algorithms: A First Course, Wiley, New York, 1985.
  • Johnson, D.W., Johnson, F.Pl, Joining Together: Group Theory and Group Skills, 5th ed., Allyn and Bacon, Boston, 1994.
  • Thomas, G.B., Finney, R.L., Calculus and Analytic Geometry, 5th ed., Addison-Wesley, Reading, 1980.
  • Whitehouse, Welchsler, Applied Operations Research: A Survey. Wiley, New York, 1976.

Return to Information Systems Links