Modeling & Simulation: Simulation & Optimization
Course: Modeling & Simulation - EEN14253
Dr. Jagat Jyoti Rath
Department of Electrical Engineering, MNNIT, Allahabad
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 1 / 15
What is Simulation Optimization?
Simulation via optimization
Difficulty and meaning
Robust heuristics
Random search
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 2 / 15
What is Optimization?
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 3 / 15
Optimization: Some examples
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 4 / 15
Optimization: Mathematical Formulation
Optimization is about solving problems of the form:
where
x is called the optimization variable or decision variable.
f(x) is called the objective function or cost function.
gi (x) are the inequality constraints.
hj (x) are the equality constraints.
In the event, the decision variables x ∈ R then its a continuous optimization and if x inZ or x is binary
then its discrete optimization.
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 5 / 15
Optimization:Solution
Solving an optimization problem usually includes the following steps:
Modeling the problem
Understanding the problem’s characteristics
Picking a numerical solver
Checking and interpreting the result
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 6 / 15
Simulation and Optimization
When the mathematical model of a system is studied using simulation, it is called a simulation
model.
A simulation experiment can be defined as a test or a series of tests in which meaningful changes are
made to the input variables of a simulation model so that we may observe and identify the reasons
for changes in the output variable(s).
The process of finding the best input variable values from among all possibilities without explicitly
evaluating each possibility is simulation optimization.
The objective of simulation optimization is minimizing the resources spent while maximizing the
information obtained in a simulation experiment.
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 7 / 15
Challenges in simulation optimization: Randomness
So why not replacing all random variables with estimates of their means?
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 8 / 15
The flaw of averages
Newsvendor Problem
The problem involves a decision-maker (the newsboy) who must decide how many newspapers to stock at
the beginning of the day to maximize his expected profit, considering that demand is uncertain.
The main steps involve the following:
Problem Setup : Decision Variables, Randomness in System, Involved costs
Formulating the profit function
Optimization approach
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 9 / 15
The flaw of averages
Newsvendor Problem
The problem involves a decision-maker (the newsboy) who must decide how many newspapers to stock at
the beginning of the day to maximize his expected profit, considering that demand is uncertain.
The main steps involve the following:
Problem Setup : Decision Variables, Randomness in System, Involved costs
Formulating the profit function
Optimization approach
What if we start using the Average Demand?
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 9 / 15
Challenges in Simulation Optimization
High computational cost
1 Simulations are often computationally expensive, especially when evaluating large-scale or
high-fidelity models.
Noisy and Stochastic Responses
1 Many simulations involve randomness, leading to noisy objective functions.
High-Dimensional Search Space
1 Many real-world problems have a large number of decision variables and constraints
Multi-Objective Complexity
1 Many simulations require optimizing multiple conflicting objectives (e.g., cost vs. performance vs.
sustainability).
Scalability and Parallelization
Large-scale simulations may require distributed computing or parallel processing.
Data and Model Uncertainty
1 Inaccurate input data or assumptions can lead to unreliable optimization results.
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 10 / 15
Formulating a Simulation Optimization Problem
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 11 / 15
Applications of Simulation Optimization
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 12 / 15
Local vs. Global Solutions
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 13 / 15
Decision Variables
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 14 / 15
Optimizing in the Presence of Noise
Dr. Jagat Jyoti Rath Modeling & Simulation: Simulation & Optimization 15 / 15