Mathematical Modeling

Mathematical Modeling
OUTLINE
Introduction
What is a mathematical model?
Why do we build a mathematical
model?
How to build a mathematical model?
An illustrative case (Case of LOD)
Formal Approaches for finding the
optimal solution
INTRODUCTION
We use the OR/MS Methodology
To capture the relationships between
various elements of the relevant
system in a mathematical model and
explore its solution.
What is a mathematical
model?
A mathematical model: Express, in
quantitative term, the relationships
between various components, as
defined in the relevant system for the
problem (e.g. using Influence Diagram).
Terminology:
Decision variables or the alternative
courses of action (controllable inputs)
Performance measure (how well the
objectives are achieved)
What is a mathematical
model?
Terminology:
Objective function (the performance
measure is expressed as a function of
decision variables)
Uncontrollable inputs: parameters,
coefficients, or constants
Constraints limit the range of the
decision variables
Relationship Between Input-SystemOutput
Why build mathematical

models?
Real-life tests are not possible
Disruptive
Risky
Expensive
Math Models are easy to

manipulate
Quick exploration of the effect of

changes in the inputs on the objective
functions
Properties of Good mathematical

models
Simple simple models are more easily
understood by the problem owner
Complete should include all significant
aspect of the problem situation affecting
the measure of effectiveness
Easy to manipulate possible to
obtain answer from the model
Adaptive changes in the structure of
the problem situation
Properties of Good mathematical

models
Easy to communicate with easy to
prepare, update, and change the inputs
and get answer quickly
Appropriate for the situation studied
produces the relevant outputs at the
lowest possible cost and in the time frame
required
Produce information that is relevant and
appropriate for decision making has to
be useful for decision making
The Art of Modeling
A scientific process
More akin to art than science
A few guidelines
Ockhams Razor:
Things should not be multiplied
without good reason.
The modeler has to be selective in
including aspects into a model
The Art of Modeling

An iterative process of enhancements
begin with a very simple model and
move in an evolutionary fashion toward
more elaborate models
Working out a numerical example
observe how variables of interest behave
Diagram and Graphs to see things in
the form of graphs or other drawings
expressing relationships and patterns.
s2-math-model-bpi-2002
Math. Model For The LOD Problem

Simplification
Constraints (Warehouse space & mixing and filling
capacities)
Two decision variables (cutoff point, L and order
size,Q)
First Approximation
Ignore the constraints
Involve only one decision variable, Q
Performance measure
Total annual relevant cost (TAC) (per year)
TAC=Annual stock holding cost+Annual set up
cost+Annual handling cost+Annual product values

Annual stock holding cost
(Average stock level x Unit product value) x Holding
cost/$/year
Annual set up cost

Setup cost per batch x Annual number of stock
replenishments
Annual handling cost

Product handling cost per unit x annual volume met
from stock
Annual product values

Unit product value x Annual volume of demand
T (Q) [0.5Qvr ] [ sD1 / Q] [h1 D1 ] [vD1 ]

T (Q, L) [ sN ] [h2 D2 ] [0.5Qvr sD1 / Q] [h1 D1 ]
T (Q L)
Math. Model LOD

[Second Approximation]
Two decision variables, L and Q.

Two additional costs
The annual set up cost for special production run
Annual volume by special prod.runs x Product
handling cost per unit
The annual handling cost for big order

Production setup per batch x Annual number of
special prod.runs
Total cost =
The annual set up cost for special production run +
The annual handling cost for big order +
Associated annual EOQ cost given L +
The annual handling cost for small order.
Deriving A Solution To The Model
Enumeration
Search Methods
Algorithmic Solution Methods
Classical Methods of Calculus
Heuristic Solution Methods
Simulation

Enumeration
Number of alternatives of action is relatively small.
Computational effort is relatively minor
Optimal solution is obtained by evaluating the
performance measure for each alternatives.
Search Methods
e.g. Golden section search
Algorithmic Solution Methods

A set of logical and mathematical operations
performed repeatedly in a specific sequence
Iteration
Stopping rules.

Classical Methods of Calculus
Heuristic Solution Methods
Impossible to find the optimal solution with the
computational means currently available (intractable)
If the optimal solution is possible to obtain, but the
potential benefit do not justify the computational effort
needed.
Heuristic methods: to find a good solutions or to improve
an existing solutions (out put based techniques)
Simulation
For complex dynamic systems
To identify good policies rather than the optimal one.

Mathematical Modeling

Uploaded by

Copyright:

Available Formats

Mathematical Modeling

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mathematical Modeling

Uploaded by

Copyright:

Available Formats

Mathematical Modeling

Relationship Between Input-SystemOutput

Why build mathematical

Math Models are easy to

Quick exploration of the effect of

Properties of Good mathematical

Properties of Good mathematical

The Art of Modeling

The Art of Modeling

Math. Model For The LOD Problem

Math. Model For The LOD Problem

Annual set up cost

Annual handling cost

Annual product values

Math. Model For The LOD Problem

T (Q) [0.5Qvr ] [ sD1 / Q] [h1 D1 ] [vD1 ]

Math. Model LOD

Two decision variables, L and Q.

The annual handling cost for big order

Deriving A Solution To The Model

Deriving A Solution To The Model

Algorithmic Solution Methods

Deriving A Solution To The Model

You might also like