BA 450

Decision Support Systems

Decision Making
Decision Making implies the existence of a
minimum of a following four factors :
Chapter II: Decision Making, Modeling, 1. There must be a problem.
and Support 2. There must be a decision maker.
3. There must be a need to solve the
problem.
4. There must be alternative solutions to
Salah BENABDALLAH, PhD
Fall Semester, 2023-2024 the problem.

1 2

Decisions Decisions…

 The managers or decision makers rely on From my experience in industry and R&D, I
decision models to come up with the best define the decision as a set of knowledge
decisions. based on analytical techniques to help
decision maker choose among a set of
 These models rely on decision theory. alternatives in light of their possible
consequences.

3 4

1
 Decisions… Models

The key word in this definition is To make a decision on complex problems,

knowledge. Therefore, for this definition to the DM may develop analytical models.
be valid we suppose that the procedure of
knowledge acquisition is active; that is the
learning process is continuous on time. We
refer by knowledge acquisition in this case
is to acquire knowledge from all available
sources.

5 6

Models… Models…

The complexity is related to three reasons: The models may be classified into three
 lack of perfect information about distinct categories based on the structure
elements and system interactions; and function of the models. They are:
 uncertainty about the nature of systems
inputs; and  Descriptive(Simulation) models
 conflicting and/or changing goals and  Prescription(Optimization) models
objectives on the part of one or more  Reasoning models
decision makers.
7 8

2
 Models… Models…
Descriptive (Simulation) models: these are Prescription (Optimization) models: these
models that attempt to describe important models are structured to capture
features of the system and the interactions important aspects of the problem
among those features. These models allow environment, but are not intended to show
the decision-maker to "test" ideas and dynamically how the system functions.
strategies and to watch what happens to These models are solved and the resulting
the system model. By testing different solution represents a policy or decision
policies, one might converge on a "good" strategy that is "optimal" with respect to
decision strategy. the assumptions that were made during the
construction of the model.
9 10

Models… Models…
Reasoning models (intelligent knowledge
based systems): these models are built on These models are built to aid a decision
the premise that machines can be maker in making decisions related to
programmed to make decisions which, if a unstructured problems or to aid a non-
human consistently made the same expert who must make decisions that are
decisions, it would be concluded that person better made by an expert.
was an expert or had special knowledge
about the domain (or area) of interest that
gave him/her key insights as to its structure
or function.
11 12

3
 The Benefits of Models The Benefits of Models…
 Mathematical models enable the analysis
 Manipulating a model is much easier of a very large, sometimes infinite, number
than manipulating a real system. of possible solutions.
 Models enable the compression of time.  The cost of making mistakes during a
 The cost of modeling analysis is much trial-and-error experiment is much lower
lower than the cost of a similar when models are used than real systems.
experiment conducted on a real system.  Models enhance and reinforce learning
 The business environment involves and training.
considerable uncertainty.
13 14