UNIT 5

Modeling

Models are abstractions of reality. Representation of an

idea or condition
Models should be valid, credible feasible and useful
Modeling is a series of steps taken to convert an idea
first into a conceptual model and then into a
quantitative model.
 Elements of modeling

Problem and objectives (purpose of the model and

simulation project)
Inputs – elements that can be altered to create an
improvement (experimental factors)
Outputs/responses (results of the simulation model
Content – components in the model and their interactions
Simplifications – ways or reducing complexity of the
problem
Assumptions – uncertainties or beliefs about real world
being modeled
 Assumptions in modeling
Models are characterized by assumptions
Variables (things which change)
Parameters (things which do not change)
Functional forms (the relationship between the two)
CLASSIFICATION OF MATHEMATICAL MODELS
Mathematical modeling process
Classification of mathematical models

Static vs. dynamic models

Lumped vs. Distributed models
Black-box vs. State space models
Linear vs. Non-linear models
Deterministic vs. Probabilistic models
Continuous vs. Discrete time models
Static vs. Dynamic models
 MODEL FOR A PREDICTION AND
ITS LIMITATIONS
PREDICTION MODEL

Predictive model is a statistical analysis of past behavior to

simulate future results
Predictive models provide quantitative estimates of an
outcome
 IMPORTANCE OF PREDICTION
MODELS IN RESEARCH STUDIES
Data mining
o Native Bayesian classifier
Natural Language Processing
o Semantic parsing, co-reference resolution and machine
translation
Medical diagnosis
o Identification of bacterial infections
Image processing
Sensor threat identification
Predicting stock market behavior
Forecasting peak bloom period for blossoms
LIMITATIONS OF PREDICTION MODELS

Lack of knowledge of initial conditions

Sensitivity to initial conditions
Occurrence of external forcing
Omission of important processes
Model imperfection
Inapplicability of safety factor
Unresolved heterogeneity
 SYSTEM SIMULATION AND
VALIDATION
Simulation

 Simulation – a technique that imitates operation of

a real-world system as it evolves over time
 Simulation is used in DSS (Decision Supporting
System)
STEPS FOLLOWED IN SIMULATION
STUDY
 Simulation in Research Studies
Simulation labs helps in medical, nursing, pharmacy
and dentistry fields
Product development program planning
Nuclear plant simulation
Reconciliation of business and systems modeling
Remote automatic incident detection
Ground – water management
Military requirements for non-war-fighting operations
 Disadvantages of simulation

Stochastic simulations produce only estimates – with

noise
Expensive to develop
Produce large volumes of output – need to summarize,
statistically analyze appropriately
Hidden critical assumptions may cause the model to
diverge from reality
Inadequate design and analysis of simulation
experiments
DYNAMIC MODELLING SYSTEM
The various stages involved in the modelling of a
Dynamic system
Physical system of interest
Modeller’s perception of the system
Mathematical representation
Manipulation to set the response
Analysis of programme

Note :
If the performance evaluation reveals that the system is not
displaying desired response, the system and its model
should be modified (or) components of the existing model
adjusted to the desired output
BASIC REQUIREMENTS FOR EMPLOYING
OPTIMIZATION TECHNIQUES
(1) Decision Variables and their relationship
The decision variables refers to any activity (product,
project, etc.) that is competing with other activities for
limited resources. The relation between variables should be
linear
(2) Well defined objective Function
A clearly well defined objective function must be stated
which may be either to maximise contribution by utilising
the available resources or it may be to produce at the lowest
cost possible by using a limited amount of productive factor
 Cont ..
(3) Presence of constraints or restrictions
There must be limitations on resources (like production,
capacity, manpower, time, machines) which are to be
allocated among various comptetive activities
(4) Alternative courses of action : It must be possible to
make a selection between various combination of the
productive factors such as men, machines, materials etc.,
(5) Non-negative restrictions : Decision variables will be
non-
negative
(6) Linearity
 Optimization
(1) Development of a production schedule by minimizing total
production and inventory costs to satisfy future demand for
a companies product
(2) Selection of a product mix which maximises the profits of
the firm, subject to several production material, marketing,
personal and financial constraints.
(3) Allocation of a limited advertising budget among various
advertising media like TV, Radio and Newspapers in order
to maximise advertising effectiveness
(4) Selection of an investment portfolio from a variety of
stocks and funds available in such a way as to maximise the
return on investment.