Multi-Objective Optimization

Multi-Objective Optimization

• Also known as multi-criteria or multi-objective decision making.

• It is a process of simultaneously optimizing two or more conflicting

objectives subject to certain constraints.

• Multi-Objective Optimization Problems (MOOP)

Multi-Objective Optimization Problems (MOOP)

• Involve more than one objective function that are to be minimized or

maximized

• Answer is set of solutions that define the best tradeoff between competing

objectives
Multi-Objective Optimization Problems (MOOP)
There are three components in any optimization problem:

F: Objectives
minimize (maximize) fi (x1, x2,· · · ,xn), i = 1,2,· · · ,m

S: Constraints
Subject to
gj (x1, x2,· · · ,xn), ROPj Cj , j = 1,2,· · · ,l

V: Design variables

xk ROPk dk , k = 1,2,· · · ,n
Multi-Objective Optimization Problems (MOOP)
Project Scheduling
Objectives:
•Minimize total project duration (makes clients happy).
•Minimize total cost (labor, equipment, materials).
•Maximize resource utilization (avoid idle time).

Use MOO to schedule activities while balancing the trade-off

between finishing early (which might require overtime pay) and
keeping costs down.
Multi-Objective Optimization Problems (MOOP)

Budget vs. Quality vs. Time

Objectives:
•Minimize construction costs.
•Maximize quality of construction (e.g., fewer defects, higher material
standards).
•Minimize project time.

Choosing between faster but expensive contractors vs. slower, cheaper ones.
Multi-Objective Optimization Problems (MOOP)
Equipment Allocation

Objectives:
•Minimize operational cost of equipment.
•Maximize equipment productivity.
•Minimize idle time.

Allocating cranes and earthmovers to different tasks on-site for maximum

efficiency with minimum overlapping.
Multi-Objective Optimization Problems (MOOP)
Labor Allocation
Objectives:
•Minimize labor costs.
•Maximize labor productivity.
•Maximize safety (fewer accidents, compliance with standards).

Example: Assigning teams to tasks based on experience, availability, and skill level.
Multi-Objective Optimization Problems (MOOP)
Let us consider, without loss of generality, a multi-objective optimization problem
with n decision variables and m objective functions

Minimize y = f (x ) = [y1 ∈ f1(x ),y2 ∈ f2(x ), · · · ,yk ∈ fm(x )]

where

x = [x1,x2,· · · ,xn] ∈ X y = [y1,y2,· · ·

,yn] ∈ Y

Here :
x is called decision vector y is called an objective vector
X is called a decision space Y is called an objective space
Decision Space and Objective Space

3 2

1
1

Thus, solving a MOOP implies to search for x in the decision space (X ) for
an optimum vector (y ) in the objective space (Y ).
Decision Space and Objective Space

3 2

1
1

The objective space is

The decision space the set of all possible
consists of all possible outcomes of the
combinations of the objective functions.
decision variables.
Multi-Objective Optimization Problems (MOOP)
We wish to determine X̄ ∈ X (called feasible region in X ) and any
point x̄ ∈ X̄ (which satisfy all the constraints in MOOP) is called
feasible solution.

Also, we wish to determine from among the set X̄ , a particular solution

x̄ ∗ that yield the optimum values of the objective functions.
Mathematically,
∀x̄ ∈ X̄ and ∃x̄ ∗ ∈ X̄ | fi (x̄ ∗ ) ≤ fi (x̄ ),

where ∀i ∈ [1,2,· · · ,m]

If this is the case, then we say that x¯∗ is a desirable solution.
Pareto Optimal Solution

Non-dominated solution set

• Given a set of solutions, the non-dominated solution set is a set of all the
solutions that are not dominated by any member of the solution set
• The non-dominated set of the entire feasible decision space is called the
Pareto-optimal set
• The boundary defined by the set of all point mapped from the Pareto optimal
set is called the Pareto optimal front
Graphical Depiction of Pareto Optimal Solution
Multi-Objective Optimization Problems (MOOP)

Three major aspects:

1. Plan Generation
2. Plan Selection
3. Plan Generation and Selection
Multi-Objective Optimization Problems (MOOP)

Plan Generation:
Plan generation refers to the process of producing a set of
feasible solutions (or "plans") that help decision-makers choose
among different trade-offs between conflicting objectives.
Plan generation is the step where multiple alternative solutions
(each representing a plan) are generated, typically forming the
Pareto front. These plans provide options for the decision-maker
to evaluate and select from, based on preferences or external
constraints.
Plan Generation Methods

1. Weightage method
2. Constraint Method
Weightage Method
Weighted Sum Method
Weighted Sum Method (Convex Case)
Weighted Sum Method (Non-Convex Case)
Weighted Sum Method

Two objective functions to minimize:

Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Weighted Sum Method
Constraint Method
Constraint Method
Constraint Method
Constraint Method
Constraint Method
Constraint Method
Constraint Method
What is Genetic Algorithm

A Genetic Algorithm (GA) is a way for a computer to

"guess" a good solution to a problem — by trying a bunch of
ideas, keeping the best ones, and mixing them to make even
better ideas.
Genetic Algorithm

• A genetic algorithm (GA) is a programming technique that

mimics biological evolution as a problem-solving strategy.
• Given a specific problem to solve, the input to the GA is a
set of potential solutions to that problem, encoded in some
fashion, and a metric called a fitness function that allows
each candidate to be quantitatively evaluated.
• These candidates may be solutions already known to work,
with the aim of the GA being to improve them, but more
often they are generated at random.
Genetic Algorithm
Genetic Algorithm
Genetic Algorithm
Genetic Algorithm
Genetic Algorithm
Multiobjective Genetic Algorithm

A Multiobjective Genetic Algorithm is a type of algorithm

designed to optimize multiple conflicting objectives
simultaneously. Rather than finding a single best solution,
MOGAs aim to find a set of Pareto optimal solutions —
offering a diverse range of trade-offs among objectives.
Key aspects of MOGA
Population-based
MOGAs work with a population of potential solutions, allowing them
to explore a wider range of the solution space compared to single-
solution algorithms
Fitness evaluation
MOGAs evaluate individuals based on their performance across
multiple objective functions, rather than a single objective function
value.
Pareto dominance
MOGAs use the concept of Pareto dominance to identify non-
dominated solutions. A solution is considered Pareto-dominated if
another solution exists that performs better on at least one objective
without performing worse on any other.
Key aspects of MOGA
Non-dominated sorting
The algorithm sorts the population into non-dominated fronts, with
higher-ranked fronts containing solutions that are less dominated.
Crowding distance
Some MOGAs, like NSGA-II, use a crowding distance metric to
maintain diversity within the non-dominated solutions and avoid
clustering of solutions.
Elite preservation
Some MOGAs include elitism, where the best solutions are directly
passed on to the next generation, helping to preserve the best-
performing solutions.
How MOGA work
Initialization: A random population of potential solutions is created.
Evaluation: The fitness of each individual is evaluated based on its performance across
multiple objective functions.
Non-dominated sorting: The population is sorted into non-dominated fronts.
Selection: Individuals are selected based on their rank and crowding distance for mating
Crossover: Genetic material is exchanged between selected individuals to create offspring
Mutation: Random changes are introduced to the offspring to explore the solution space
Replacement: The new offspring population replaces the old population
Iteration: Steps 2-7 are repeated until a stopping criterion is met (e.g., a maximum number
of iterations, or a satisfactory Pareto front has been found.
Multiobjective Genetic Algorithm
Multiobjective Genetic Algorithm: Example
Multiobjective Genetic Algorithm: Example
Multiobjective Genetic Algorithm: Example
Designing a new car and to optimize for fuel efficiency, performance,
and cost. These objectives are often conflicting - maximizing
performance might decrease fuel efficiency, and a high-performance
car might be more expensive. A MOGA can help to find a set of car
designs that represent different trade-offs between these objectives,
allowing you to choose the design that best fits your specific needs and
priorities.
Multiobjective Genetic Algorithm: Example

1.1. Population Initialization: A population of candidate solutions (representing

different car designs) is randomly generated.

2.2. Evaluation: Each solution is evaluated based on its performance across all
objectives (fuel efficiency, performance, and cost).

3.3. Selection: Solutions are selected for reproduction based on their

performance. Better solutions are more likely to be selected.
4.4 Crossover: Selected solutions (parents) combine their genes to create offspring
solutions.
5.5 Mutation: Small random changes (mutations) are introduced to the offspring,
which can lead to new and potentially better solutions.
Multiobjective Genetic Algorithm: Example

Non-Dominated Sorting:
Solutions are sorted based on whether they are dominated by other solutions. A
solution is dominated if there exists another solution that performs at least as well
in all objectives and better in at least one objective.
Pareto Front:
The non-dominated solutions form the Pareto front, which represents the optimal
trade-off between the objectives.
Iteration:
Steps 2-7 are repeated for several generations until a satisfactory Pareto front is
found.
Question Bank

1. Define plan generation in the context of construction project planning. What

are the key inputs and expected outputs?
2. Explain the application of weightage optimization method with suitable
example of construction planning.
3. What is the weightage method in multi-criteria decision-making? Illustrate with
an example from contractor selection or project procurement evaluation.
4. Discuss the pros and cons of using the weightage method in evaluating project
alternatives such as materials selection, vendor choice, or site layout planning.
5. Discuss how traditional construction planning differs from computational
planning
Question Bank

1. Explain the concept of Multi-objective Genetic Algorithms. How can it be

applied to optimize a construction schedule that minimizes cost and maximizes
resource utilization?
2. What are the Multiobjective Genetic Algorithms? Explain the Process of
Multiobjective Genetic Algorithms with suitable example using flowchart.