Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 16

INTRODUCTION

BIG DATA OPTIMIZATION


20CS3276RB
LTPS-2020
Motivation
• Currently, we are in the Information Age, where more
organizational and individual activities and processes are
based on Information Technology. We are also in a fast
changing world.

• Due to several factors, such as globalization, technological


improvements, and more recently the 2008 financial crisis,
both organizations and individuals are pressured for improving
their efficiency, reducing costs, and making better-informed
decisions.

• This is where optimization methods, supported by


computational tools, can play a key role.
Optimization and its applications
• Any real world problem can be viewed as an optimization
problem.

• Optimization is about minimizing or maximizing a goal (or


goals).

• It is useful in various domains like Agriculture, Banking,


Engineering, Finance, Marketing, Production and Science.

• Examples of real-world applications include the optimization


of construction works, financial portfolios, marketing
campaigns, and water management in agriculture, just to
name a few.
• Rather than taking decisions based on human experience and
intuition, there is an increasing trend for adopting computational
tools, based on optimization methods, to analyse real-world data in
order to make better informed decisions.

• Modern optimization, also known as metaheuristics.

• Modern optimization methods are general purpose solvers, i.e.,


applicable to a wide range of distinct problems, since few domain
knowledge is required.

• Modern optimization is particularly useful for solving complex


problems for which no specialized optimization algorithm has been
developed, such as problems with discontinuities, dynamic changes,
multiple objectives, or hard and soft restrictions, which are more
difficult to be handled by classical methods.
• A metaheuristic is a higher-level procedure or heuristic
designed to find, generate, or select a heuristic (partial
search algorithm) that may provide a sufficiently good
solution to an optimization problem, especially with
incomplete or imperfect information or limited
computation capacity.
• The Goal is to efficiently explore the search space in
order to find near- optimal solutions.
• Method like
 Hill climbing
 Ant colony optimization
 Genetic Algorithm
 Particle Swarm optimization
R Language
• R is a free, open source, and multiple platform tool (e.g.,
Windows, Linux, MacOS) that was specifically developed for
statistical analysis.

• The R community is very active and new packages are being


continuously created, with more than 5800 packages available,
thus enhancing the tool capabilities.

• In particular, several of these packages implement modern


optimization methods.
Two types of practitioners of R
• R users interested in applying modern optimization methods
and non-R expert data analysts or optimization practitioners
who want to test the R capabilities for optimizing real-world
tasks.
Representation of a Solution
• A major decision when using modern optimization methods is
related with how to represent a possible solution.

• Such decision sets the search space and its size, thus producing
an impact on how new solutions are searched.

• To represent a solution, there are several possibilities. Binary,


integer, character, real value and ordered vectors, matrices,
trees and virtually any computer based representation form
(e.g., computer program) can be used to encode solutions.
EXAMPLES
Evaluation Function
• Another important decision for handling optimization tasks is
the definition of the evaluation function, which should translate
the desired goal (or goals) to be maximized or minimized.

• Such function allows to compare different solutions, by


providing either a rank or a quality measure score.

• When considering numeric functions, the shape can be convex


or non-convex, with several local minima/maxima.
Constraints
• There are two main types of constraints:
hard and soft

• Hard constraints cannot be violated and are due to factors


such as laws or physical restrictions.

• Soft constraints are related with other (often non-priority)


user goals, such as increasing production efficiency while
reducing environmental costs.
Optimization Methods
• There are different dimensions that can be used to classify
optimization methods.

• Three types of analysis adopted are:


1. The type of guided search
2. The search is deterministic or stochastic based;
3. and if the method is inspired by physical or biological processes.
DEMONSTRATIVE PROBLEMS

1. sum of bits:

2. max sin:

You might also like