Differential Evolution Hossein Talebi Hassan nikoo 1.

Slides:



Advertisements
Similar presentations
Particle Swarm Optimization (PSO)
Advertisements

Population-based metaheuristics Nature-inspired Initialize a population A new population of solutions is generated Integrate the new population into the.
Constraint Optimization We are interested in the general non-linear programming problem like the following Find x which optimizes f(x) subject to gi(x)
Particle Swarm Optimization
Parameter Control A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Chapter 8.
CHAPTER 9 E VOLUTIONARY C OMPUTATION I : G ENETIC A LGORITHMS Organization of chapter in ISSO –Introduction and history –Coding of  –Standard GA operations.
1 Lecture 8: Genetic Algorithms Contents : Miming nature The steps of the algorithm –Coosing parents –Reproduction –Mutation Deeper in GA –Stochastic Universal.
Effective gradient-free methods for inverse problems Jyri Leskinen FiDiPro DESIGN project.
Genetic Algorithms in Materials Processing N. Chakraborti Department of Metallurgical & Materials Engineering Indian Institute of Technology Kharagpur.
Intro to AI Genetic Algorithm Ruth Bergman Fall 2002.
Evolutionary Computational Intelligence Lecture 8: Memetic Algorithms Ferrante Neri University of Jyväskylä.
D Nagesh Kumar, IIScOptimization Methods: M1L4 1 Introduction and Basic Concepts Classical and Advanced Techniques for Optimization.
Intro to AI Genetic Algorithm Ruth Bergman Fall 2004.
Differential Evolution Hossein Talebi Hassan Nikoo 1.
Optimization of thermal processes2007/2008 Optimization of thermal processes Maciej Marek Czestochowa University of Technology Institute of Thermal Machinery.
Particle Swarm Optimization Algorithms
1 Reasons for parallelization Can we make GA faster? One of the most promising choices is to use parallel implementations. The reasons for parallelization.
IE 607 Constrained Design: Using Constraints to Advantage in Adaptive Optimization in Manufacturing.
A Genetic Algorithms Approach to Feature Subset Selection Problem by Hasan Doğu TAŞKIRAN CS 550 – Machine Learning Workshop Department of Computer Engineering.
Evolutionary Intelligence
© Negnevitsky, Pearson Education, CSC 4510 – Machine Learning Dr. Mary-Angela Papalaskari Department of Computing Sciences Villanova University.
Evolutionary Algorithms 3.Differential Evolution.
Multimodal Optimization (Niching) A/Prof. Xiaodong Li School of Computer Science and IT, RMIT University Melbourne, Australia
Example II: Linear truss structure
Slides are based on Negnevitsky, Pearson Education, Lecture 12 Hybrid intelligent systems: Evolutionary neural networks and fuzzy evolutionary systems.
Introduction to Genetic Algorithms and Evolutionary Computation
A Comparison of Nature Inspired Intelligent Optimization Methods in Aerial Spray Deposition Management Lei Wu Master’s Thesis Artificial Intelligence Center.
Hierarchical Distributed Genetic Algorithm for Image Segmentation Hanchuan Peng, Fuhui Long*, Zheru Chi, and Wanshi Siu {fhlong, phc,
Genetic algorithms Prof Kang Li
Zorica Stanimirović Faculty of Mathematics, University of Belgrade
Boltzmann Machine (BM) (§6.4) Hopfield model + hidden nodes + simulated annealing BM Architecture –a set of visible nodes: nodes can be accessed from outside.
Optimization Problems - Optimization: In the real world, there are many problems (e.g. Traveling Salesman Problem, Playing Chess ) that have numerous possible.
(Particle Swarm Optimisation)
Outline Introduction Evolution Strategies Genetic Algorithm
2 Fundamentals of Genetic Algorithms Alexandre P. Alves da Silva and Djalma M. Falca˜o.
Distributed Genetic Algorithms with a New Sharing Approach in Multiobjective Optimization Problems Tomoyuki HIROYASU Mitsunori MIKI Sinya WATANABE Doshisha.
© Negnevitsky, Pearson Education, Lecture 9 Evolutionary Computation: Genetic algorithms Introduction, or can evolution be intelligent? Introduction,
15/06/2003NORPIE 2004, Trondheim1 Genetic Optimization of Electric Machines, a State of the Art Study S. E. Skaar, R. Nilssen.
2005MEE Software Engineering Lecture 11 – Optimisation Techniques.
1 Effect of Spatial Locality on An Evolutionary Algorithm for Multimodal Optimization EvoNum 2010 Ka-Chun Wong, Kwong-Sak Leung, and Man-Hon Wong Department.
Evolution strategies Chapter 4. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Evolution Strategies ES quick overview Developed: Germany.
Biologically inspired algorithms BY: Andy Garrett YE Ziyu.
CITS7212: Computational Intelligence An Overview of Core CI Technologies Lyndon While.
Parallel Genetic Algorithms By Larry Hale and Trevor McCasland.
Evolution strategies Chapter 4. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Evolution Strategies ES quick overview Developed: Germany.
A survey of Constraint Handling Techniques in Evolutionary Computation Methods Author: Zbigneiw Michalewicz Presenter: Masoud Mazloom 27 th Oct
D Nagesh Kumar, IIScOptimization Methods: M8L5 1 Advanced Topics in Optimization Evolutionary Algorithms for Optimization and Search.
Multi-objective Evolutionary Algorithms (for NACST/Seq) summarized by Shin, Soo-Yong.
1 Autonomic Computer Systems Evolutionary Computation Pascal Paysan.
Evolutionary multi-objective algorithm design issues Karthik Sindhya, PhD Postdoctoral Researcher Industrial Optimization Group Department of Mathematical.
Application of the GA-PSO with the Fuzzy controller to the robot soccer Department of Electrical Engineering, Southern Taiwan University, Tainan, R.O.C.
Genetic Algorithm Dr. Md. Al-amin Bhuiyan Professor, Dept. of CSE Jahangirnagar University.
Artificial Intelligence By Mr. Ejaz CIIT Sahiwal Evolutionary Computation.
Genetic Algorithms. Solution Search in Problem Space.
Breeding Swarms: A GA/PSO Hybrid 簡明昌 Author and Source Author: Matthew Settles and Terence Soule Source: GECCO 2005, p How to get: (\\nclab.csie.nctu.edu.tw\Repository\Journals-
Constrained Optimization by the  Constrained Differential Evolution with an Archive and Gradient-Based Mutation Tetsuyuki TAKAHAMA ( Hiroshima City University.
Genetic Algorithm. Outline Motivation Genetic algorithms An illustrative example Hypothesis space search.
 Negnevitsky, Pearson Education, Lecture 12 Hybrid intelligent systems: Evolutionary neural networks and fuzzy evolutionary systems n Introduction.
ZEIT4700 – S1, 2016 Mathematical Modeling and Optimization School of Engineering and Information Technology.
Evolutionary Algorithms Jim Whitehead
Particle Swarm Optimization (2)
C.-S. Shieh, EC, KUAS, Taiwan
Probability-based Evolutionary Algorithms
Meta-Heuristic Algorithms 16B1NCI637
Multi-Objective Optimization
○ Hisashi Shimosaka (Doshisha University)
EE368 Soft Computing Genetic Algorithms.
Boltzmann Machine (BM) (§6.4)
Population Methods.
Stochastic Methods.
Presentation transcript:

Differential Evolution Hossein Talebi Hassan nikoo 1

Variations to Basic Differential Evolution  Hybrid Differential Evolution Strategies  Gradient-Based Hybrid Differential Evolution  Evolutionary Algorithm-Based Hybrids  DE Reproduction Process used as Cross over Operation of Simple GA  Ranked-Based Cross Over Operator for DE  Particle Swarm Optimization Hybrids  Population-Based Differential Evolution  Self-Adaptive Differential Evolution 2

Variations to Basic Differential Evolution  Differential Evolution for Discrete-Valued Problems  Angle Modulated Differential Evolution  Binary Differential Evolution  Constraint Handling Approaches  Multi-Objective Optimization  Dynamic Environments  Applications 3

Gradient-Based Hybrid Differential Evolution  acceleration operator to improve convergence speed – without decreasing diversity  migration operator  The Acceleration Operator uses gradient descent to adjust the best individual toward obtaining a better position 4

Gradient-Based Hybrid Differential Evolution(Cont.)  Using Gradient Descent may results in getting stuck in a local optima or premature convergence  We can increase Population diversity by Migration Operator  This operator spawns new individuals from the best individual and replaces the current population with these new individuals 5

Gradient-Based Hybrid Differential Evolution(Cont.)  The Migration Operator is applied only when diversity of population becomes too small 6

Gradient-Based Hybrid Differential Evolution(Cont.) 7

Using Stochastic Gradient Descent and DE for Neural Networks Training  Stochastic Gradient Descent 8

Evolutionary Algorithm-Based Hybrids  Hrstka and Kucerov´a used the DE reproduction process as a crossover operator in a simple GA  Chang and Chang used standard mutation operators to increase DE population diversity by adding noise to the created trial vectors. 9

Evolutionary Algorithm-Based Hybrids(Cont.)  Sarimveis and Nikolakopoulos [758] use rank-based selection to decide which individuals will take part to calculate difference vectors 10

 Ranked-base Mutation 11

Particle Swarm Optimization Hybrids  Hendtlass proposed that the DE reproduction process be applied to the particles in a PSO swarm at specified intervals.  Kannan et al. apply DE to each particle for a number of iterations and replaces the best with particle 12

Particle Swarm Optimization Hybrids(Cont.)  Another approach is to change only change best particle using  Where sigma is general difference vector 13

Population-Based Differential Evolution  Ali and T¨orn proposed to use an auxiliary population  For each offspring created, if the fitness of the offspring is not better than the parent, instead of discarding the offspring, it is considered for inclusion in the auxiliary Population 14

DETVSF (DE with Time Varying Scale Factor)  During the later stages it is important to adjust the movements of trial solutions finely so that they can explore the interior of a relatively small space in which the suspected global optimum lies  We can reduce the scale factor linearly with time from a (predetermined) maximum to a (predetermined) minimum value 15

DETVSF (DE with Time Varying Scale Factor)(Cont.) 16

Parameter Control in DE  Dynamic Parameters  Self-Adaptive 17

Self-Adaptive Parameters  probability of recombination be self – adapted  Mu is the average of successful probablities  Abbass Proposed to use this formula : 18

Self-Adaptive Parameters(Cont.)  Omran et al. propose a self-adaptive DE strategy that makes use of this formula for scale factor  For mutation operator 19

Angle Modulated Differential Evolution  Pampar´a et al. proposed a DE algorithm to evolve solutions to binary-valued optimization problems, without having to change the operation of the original DE  They use a mapping between binary- valued and continuous-valued space to solve the problem in binary space 20

Angle Modulated Differential Evolution(Cont.)  The objective is to evolve, in the abstracted continues space, a bitstring generating function will be used in the original space to produce bit-vector solutions  ‘a’, ’b’, ‘c’ and ‘d’ are continues space problem parameter 21

Angle Modulated Differential Evolution(Cont.)  ‘a=0’ ‘b=1’ ‘c=1’ ‘d=0’ 22

Binary Differential Evolution  binDE borrows concepts from the binary particle swarm optimizer binPSO  binDE uses the floating-point DE individuals to determine a probability for each component  the corresponding bitstring solution will be calculated as follow : 23

Binary Differential Evolution 24

Constraint Handling Approaches  Penalty methods  adding a function to penalize solutions that violate constraints  Using F(x, t) = f(x, t) + λp(x, t) where λ is the penalty coefficient and p is time dependent penalty function  Converting the constrained problem to an unconstrained problem 25

Constraint Handling Approaches(Cont.)  We can convert constrained problem to an unconstrained problem by defining the Lagrangian for the constrained problem  If primal problem is convex then defining dual problem and solving minmax problem 26

Constraint Handling Approaches(Cont.)  By changing selection operator, infeasible solutions can be rejected and we can use a method for repairing of the infeasible solution 27

Constraint Handling Approaches(Cont.)  Boundary constraints are easily enforced by clamping offspring to remain within the given boundaries 28

Multi-Objective Optimization  Converting the problem into the  Weighted Aggregation Methods 29

Multi-Objective Optimization(Cont.)  This method intends to define an aggregate objective function as a weighted sum of the objectives  Usually assumed that 30

Multi-Objective Optimization(Cont.)  There is no guarantee that different solutions will be found  A niching strategy can be used to find multiple solutions  It is difficult to get the best weight values, ωk, since these are problem-dependent 31

Multi-Objective Optimization(Cont.)  Vector evaluated DE is a population based method for MOO  If K objectives have to be optimized, K sub-populations are used, where each subpopulation optimizes one of the objectives.  Sub-populations are organized in a ring topology  The best individual of sub-population Ck migrates to population Ck+1 to produce the trial vectors for that population 32

Dynamic Environments  Assumptions  the number of peaks, n X, to be found are know and these peaks are evenly distributed through the search space  Changes are small and gradual  DynDE uses multiple populations, with each population maintaining one of the peaks 33

Dynamic Environments(Cont.)  At each iteration, the best individuals of each pair of sub-populations are compared if these global best positions are too close to one another, the sub- population with the worst global best solution is re-initialized 34

Dynamic Environments(Cont.)  The following diversity increasing strategies  Re-initialize the sub-populations  Use quantum individuals :Some of the individuals are re-initialized to random points inside a ball centered at the global best individual  Use Brownian individuals: Some positions are initialized to random positions around global best individual  Some individuals are simply added noise 35

Dynamic Environments(Cont.)  Initialization of Quantum Individuals 36

Applications  Mostly applied to optimize functions defined over continuous-valued landscapes  Clustering  Controllers  Filter design  Image analysis  Integer-Programming  Model selection  NN training 37

References 1. Computational Intelligence, an introduction,2 nd edition, Andries Engelbercht, Wiley 2. Differential Evolution - A simple and efficient adaptive scheme for global optimization over continuous spaces, Rainer Storn,Kenneth Price, Particle Swarm Optimization and Differential Evolution Algorithms: Technical Analysis, Applications and Hybridization Perspectives, Swagatam Das1, Ajith Abraham2, and Amit Konar1,Springer Differential Evolution, homepage Differential Evolution 38

Thanks For Your Attention Any Question? Differential Evolution 39