Solving Permutation Problems with Estimation of Distribution Algorithms and Extensions Thereof Josu Ceberio.

Slides:



Advertisements
Similar presentations
L3S Research Center University of Hanover Germany
Advertisements

Josu Ceberio. Previously…  EDAs for integer domains.  EDAs for real value domains.  Few efficient designs for permutation- based problems. POOR PERFORMANCE.
CS6800 Advanced Theory of Computation
1 An Adaptive GA for Multi Objective Flexible Manufacturing Systems A. Younes, H. Ghenniwa, S. Areibi uoguelph.ca.
Fast Algorithms For Hierarchical Range Histogram Constructions
Multi-Objective Optimization NP-Hard Conflicting objectives – Flow shop with both minimum makespan and tardiness objective – TSP problem with minimum distance,
Algorithms + L. Grewe.
5-1 Chapter 5 Tree Searching Strategies. 5-2 Satisfiability problem Tree representation of 8 assignments. If there are n variables x 1, x 2, …,x n, then.
Optimal Design Laboratory | University of Michigan, Ann Arbor 2011 Design Preference Elicitation Using Efficient Global Optimization Yi Ren Panos Y. Papalambros.
Reporter : Mac Date : Multi-Start Method Rafael Marti.
Introduction to Evolutionary Computation  Genetic algorithms are inspired by the biological processes of reproduction and natural selection. Natural selection.
MAE 552 – Heuristic Optimization Lecture 26 April 1, 2002 Topic:Branch and Bound.
EAs for Combinatorial Optimization Problems BLG 602E.
Ant Colony Optimization Optimisation Methods. Overview.
1 Branch and Bound Searching Strategies 2 Branch-and-bound strategy 2 mechanisms: A mechanism to generate branches A mechanism to generate a bound so.
Josu Ceberio Alexander Mendiburu Jose A. Lozano
D Nagesh Kumar, IIScOptimization Methods: M1L4 1 Introduction and Basic Concepts Classical and Advanced Techniques for Optimization.
Review Rong Jin. Comparison of Different Classification Models  The goal of all classifiers Predicating class label y for an input x Estimate p(y|x)
Ant Colony Optimization: an introduction
Metaheuristics The idea: search the solution space directly. No math models, only a set of algorithmic steps, iterative method. Find a feasible solution.
Elements of the Heuristic Approach
Toshihide IBARAKI Mikio KUBO Tomoyasu MASUDA Takeaki UNO Mutsunori YAGIURA Effective Local Search Algorithms for the Vehicle Routing Problem with General.
Genetic Algorithms and Ant Colony Optimisation
Efficient Model Selection for Support Vector Machines
On comparison of different approaches to the stability radius calculation Olga Karelkina Department of Mathematics University of Turku MCDM 2011.
Methods of Solving the Problem of Construction the Optimum Regression Model as a Discrete Optimization Task Ivan Melnyk The International Research and.
1 Paper Review for ENGG6140 Memetic Algorithms By: Jin Zeng Shaun Wang School of Engineering University of Guelph Mar. 18, 2002.
FDA- A scalable evolutionary algorithm for the optimization of ADFs By Hossein Momeni.
Swarm Intelligence 虞台文.
Introduction to Job Shop Scheduling Problem Qianjun Xu Oct. 30, 2001.
Search Methods An Annotated Overview Edward Tsang.
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)
1 Exploring Custom Instruction Synthesis for Application-Specific Instruction Set Processors with Multiple Design Objectives Lin, Hai Fei, Yunsi ACM/IEEE.
Design of an Evolutionary Algorithm M&F, ch. 7 why I like this textbook and what I don’t like about it!
Guaranteed Convergence and Distribution in Evolutionary Multi- Objective Algorithms (EMOA’s) via Achivement Scalarizing Functions By Karthik.
The Application of The Improved Hybrid Ant Colony Algorithm in Vehicle Routing Optimization Problem International Conference on Future Computer and Communication,
1 Short Term Scheduling. 2  Planning horizon is short  Multiple unique jobs (tasks) with varying processing times and due dates  Multiple unique jobs.
Course: Logic Programming and Constraints
Artificial Intelligence Chapter 4. Machine Evolution.
Antonio Augusto Chaves - Luiz Antonio Nogueira Lorena National Institute for Space Research - INPE São José dos Campos, Brazil
Exact and heuristics algorithms
1 Genetic Algorithms and Ant Colony Optimisation.
Genetic Algorithms Przemyslaw Pawluk CSE 6111 Advanced Algorithm Design and Analysis
Kanpur Genetic Algorithms Laboratory IIT Kanpur 25, July 2006 (11:00 AM) Multi-Objective Dynamic Optimization using Evolutionary Algorithms by Udaya Bhaskara.
SemiBoost : Boosting for Semi-supervised Learning Pavan Kumar Mallapragada, Student Member, IEEE, Rong Jin, Member, IEEE, Anil K. Jain, Fellow, IEEE, and.
Introducing Mixtures of Generalized Mallows in Estimation of Distribution Algorithms Josian Santamaria Josu Ceberio Roberto Santana Alexander Mendiburu.
Edge Assembly Crossover
Genetic Algorithms. 2 Overview Introduction To Genetic Algorithms (GAs) GA Operators and Parameters Genetic Algorithms To Solve The Traveling Salesman.
Kernels of Mallows Models for Solving Permutation-based Problems
Diversity Loss in General Estimation of Distribution Algorithms J. L. Shapiro PPSN (Parallel Problem Solving From Nature) ’06 BISCuit 2 nd EDA Seminar.
Yuan-Ze University A Genetic Algorithm with Injecting Artificial Chromosomes for Single Machine Scheduling Problems Pei-Chann Chang, Shih-Shin Chen, Qiong-Hui.
Feature Selction for SVMs J. Weston et al., NIPS 2000 오장민 (2000/01/04) Second reference : Mark A. Holl, Correlation-based Feature Selection for Machine.
GENETIC ALGORITHMS Tanmay, Abhijit, Ameya, Saurabh.
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.
HIGH LEVEL SYNTHESIS WITH AREA CONSTRAINTS FOR FPGA DESIGNS: AN EVOLUTIONARY APPROACH Tesi di Laurea di: Christian Pilato Matr.n Relatore: Prof.
Branch and Bound Searching Strategies
Metaheuristics for the New Millennium Bruce L. Golden RH Smith School of Business University of Maryland by Presented at the University of Iowa, March.
1 Comparative Study of two Genetic Algorithms Based Task Allocation Models in Distributed Computing System Oğuzhan TAŞ 2005.
Genetic Algorithms. Solution Search in Problem Space.
Bulgarian Academy of Sciences
School of Computer Science & Engineering
Example: Applying EC to the TSP Problem
Comparing Genetic Algorithm and Guided Local Search Methods
Example: Applying EC to the TSP Problem
Artificial Intelligence Chapter 4. Machine Evolution
Multi-Objective Optimization
Artificial Intelligence Chapter 4. Machine Evolution
Survey on Coverage Problems in Wireless Sensor Networks
Presentation transcript:

Solving Permutation Problems with Estimation of Distribution Algorithms and Extensions Thereof Josu Ceberio

Outline Permutation optimization problems Part I : Contributions to the design of Estimation of Distribution Algorithms for permutation problems Part II: Studying the linear ordering problem Part III: A general multi-objectivization scheme based on the elementary landscape decomposition Conclusions and future work 2

Combinatorial optimization problems Permutation optimization problems Definition 3

Problems whose solutions are naturally represented as permutations 4

Permutation optimization problems Notation 5 A permutation is a bijection of the set onto itself,

Permutation optimization problems Goal To find the permutation solution that minimizes a fitness function The search space consists of solutions. 6

Permutation optimization problems Travelling salesman problem (TSP) Permutation Flowshop Scheduling Problem (PFSP) Linear Ordering Problem (LOP) Quadratic Assignment Problem (QAP) 7

Permutation optimization problems Travelling Salesman Problem (TSP) Which permutation of cities provides the shortest path?

Permutation optimization problems Travelling Salesman Problem (TSP) Which permutation of cities provides the shortest path?

Permutation optimization problems Travelling Salesman Problem (TSP) Possible routes:

Permutation optimization problems Definition Many of these problems are NP-hard. (Garey and Johnson 1979) 11

Contributions to the design of EDAs for permutation problems Part I

Estimation of distribution algorithms Definition 13

Review of EDAs for permutation problems EDAs for integer domain problems –The sampling step may not provide permutations, but solutions in. 14 Learn a probability distribution over the set

Review of EDAs for permutation problems EDAs for integer domain problems –The sampling step may not provide permutations, but solutions in. –The probabilistic logic sampling is modified to guarantee mutual exclusivity constraints. 15 Learn a probability distribution over the set

Review of EDAs for permutation problems EDAs for integer domain problems –The sampling step may not provide permutations, but solutions in. –The probabilistic logic sampling is modified to guarantee mutual exclusivity constraints. –EDAs that have used this approach: UMDA MIMIC EBNA TREE … 16 Learn a probability distribution over the set

Review of EDAs for permutation problems EDAs for continuous domain problems  The probability of a given permutation cannot be calculated in closed form.  Sample solutions of real values (0.30, 0.10, 0.40, 0.20) (0.27, 0.62, 0.71, 0.20) Learn a probability distribution on the continuous domain

Review of EDAs for permutation problems EDAs for continuous domain problems  Highly redundant codification (0.30, 0.10, 0.40, 0.20) (0.25, 0.14, 0.35, 0.16) (0.60, 0.20, 0.80, 0.40) (0.27, 0.15, 0.31, 0.20) (0.83, 0.01, 0.99, 0.70) (0.37, 0.07, 0.75, 0.36) (0.60, 0.50, 0.71, 0.52) (0.17, 0.05, 0.21, 0.10)  EDAs that have used this approach: UMDA c, MIMIC c, EGNA… 18 Learn a probability distribution on the continuous domain

Position Item Review of EDAs for permutation problems Permutation-oriented EDAs Node and Edge Histogram-based Sampling Algorithms (EHBSA & NHBSA) (Tsutsui et al. 2002, Tsutsui et al. 2006) 19 Node Histogram Population

Item j Item i Review of EDAs for permutation problems Permutation-oriented EDAs Node and Edge Histogram-based Sampling Algorithms (EHBSA & NHBSA) (Tsutsui et al. 2002, Tsutsui et al. 2006) 20 Edge Histogram Population

Review of EDAs for permutation problems Permutation-oriented EDAs Node and Edge Histogram-based Sampling Algorithms (EHBSA & NHBSA) (Tsutsui et al. 2002, Tsutsui et al. 2006) 21 Parent Offspring Template Strategy (WT)

967 Review of EDAs for permutation problems Permutation-oriented EDAs Node and Edge Histogram-based Sampling Algorithms (EHBSA & NHBSA) (Tsutsui et al. 2002, Tsutsui et al. 2006) Parent Offspring Sample from the model 425 Template Strategy (WT) 813

Review of EDAs for permutation problems Permutation-oriented EDAs IDEA- Induced Chromosome Elements Exchanger (ICE) (Bosman and Thierens 2001)  A continuous domain EDA hybridized with a crossover operator Recursive EDA (REDA) (Romero and Larrañaga 2009)  A k stages algorithm, where at each stage, a specific part of the individual is optimized with an EDA  UMDA, MIMIC,…. 23

Review of EDAs for permutation problems Experimental design EDAs: UMDA, MIMIC, EBNA BIC, TREE UMDA c, MIMIC c, EGNA NHBSA WT, NHBSA WO, EHBSA WT,EHBSA WO, IDEA-ICE, REDA UMDA, REDA MIMIC OmeGA. 4 problems and 100 instances (25 instances of each problem). Average of 20 repetitions of each algorithm. Statistical test: Friedman + Shaffer’s static procedure. 24

Review of EDAs for permutation problems Experiments TSP Best performing algorithms: NHBSA WT, EHBSA WT. 25 Critical difference diagram

Review of EDAs for permutation problems Experiments 26 Critical difference diagram Estimate first and second order marginal probabilities. TSP

Three research paths to investigate Learn models based on high order marginal probabilities –K-order marginals-based EDA Implement probability models for permutation domains –The Mallows EDA –The Generalized Mallows EDA –The Plackett-Luce EDA Non-parametric models  Kernels of Mallows models. 27

The Mallows model Definition A distance-based exponential probability model Central permutation Spread parameter A distance on permutations 28

The Mallows model Definition A distance-based exponential probability model Central permutation Spread parameter A distance on permutations 29

The Mallows model Definition A distance-based exponential probability model Central permutation Spread parameter A distance on permutations 30

The Generalized Mallows model Definition If the distance can be decomposed as sum of terms then, the Mallows model can be generalized as The Generalized Mallows model n-1 spread parameters 31

The Generalized Mallows model Kendall’s-τ distance 32 Kendall’s-τ distance: calculates the number of pairwise disagreements

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda Population Average solution (,,,, )

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda Population Average solution ( 2.7,,,, )

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda. 35 Population Average solution ( 2.7, 2.9,,, )

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda. 36 Population Average solution ( 2.7, 2.9, 3.2,, )

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda. 37 Population Average solution ( 2.7, 2.9, 3.2, 3.7, )

The Generalized Mallows model Learning and sampling Learning in 2 steps: Calculate the central permutation by means of Borda. 38 Population Average solution ( 2.7, 2.9, 3.2, 3.7, 2.5 )

Learning in 2 steps: Calculate the central permutation by means of Borda. Maximum likelihood estimation of the spread parameters. Upper bounds are set to avoid premature convergence. Sampling in 2 steps: Sample a vector from Build a permutation from the vector and The Generalized Mallows model Learning and sampling 39

Permutation Flowshop Scheduling Problem Definition Total flow time (TFT) m1m1 m2m2 m3m3 m4m4 j4j4 j1j1 j3j3 j2j2 j5j5 jobs machines processing times 5 x 4 40

The number of evaluations performed by AGA in n x m x 0.4s State-of-the-art algorithms: Asynchronous Genetic Algorithm (AGA) (Xu et al. 2011) Initialize with LR(n/m) (Li and Reeves 2001) Genetic algorithm with local search Variable Neighborhood Search 4 (VNS 4 ) (Costa et al. 2012) Initialize with LR(n/m) (Li and Reeves 2001) 220 instances from Taillard’s and Random benchmarks. 20 repetitions Stopping criterion Experimental design 41 Execution time: n x m x 0.4s

The Generalized Mallows EDA Experiments AGAVNS 4 GMEDAAGAVNS 4 GMEDA 20 x x x x20 20 x x x x x x x x x x x x x x x x x x instances 42

Hybrid Generalized Mallows EDA HGMEDA 43 Best solution GMEDA VNS Half evaluations

The Hybrid Generalized Mallows EDA Experiments GMEDAVNSHGMEDAGMEDAVNSHGMEDA 20 x x x x20 20 x x x x x x x x x x x x x x x x x x instances 44

The Hybrid Generalized Mallows EDA Experiments AGAVNS 4 HGMEDAAGAVNS 4 HGMEDA 20 x x x x20 20 x x x x x x x x x x x x x x x x x x instances 45

The Generalized Mallows EDA Analysis 46

The Generalized Mallows EDA Analysis 47

The Generalized Mallows EDA Analysis 48

Experimental design 49 State-of-the-art algorithms: Asynchronous Genetic Algorithm (AGA): Initialize with LR Genetic algorithm with local search Variable Neighborhood Search 4 (VNS 4 ) 220 instances from Taillard’s and Random benchmarks. 20 repetitions Stopping criterion n x m x 0.4s number of evaluations Guided HGMEDA

The Generalized Mallows EDA LR initialization and additional evaluations 50

278 instances The Generalized Mallows EDA Conclusions A new EDA that codifies a probability model for permutation domains was proposed. An algorithm based on the Generalized Mallows EDA outperformed existing state-of-the-art algorithms in 152 instances of the PFSP out of 220. The analysis pointed out that the contribution of the Generalized Mallows model has been essential in this achievement. 51

Other distances Cayley distance Calculates the minimum number of swap operations to convert in. 52

Other distances Ulam distance 53 Calculates the minimum number of insert operations to convert in.

EDAs: Mallows – Kendall (M Ken ) Mallows – Cayley (M Cay ) Mallows – Ulam (M Ula ) Generalized Mallows – Kendall (GM Ken ) Generalized Mallows – Cayley (GM Cay ) 4 problems: TSP, LOP, PFSP, QAP 50 instances for each problem of sizes: 10,20,30,40,50,60,70,80,90, repetitions Stopping criterion: 1000n 2 evaluations Experimental design 54

55 Evaluating the performance of EDAs GM cay M ula GM cay

Distances and neighborhoods –Two solutions and are neighbors if the Kendall’s-τ distance between and is 56 –Two solutions and are neighbors if the Cayley distance between and is –Two solutions and are neighbors if the Ulam distance between and is Swap neighborhood Interchange neighborhood Insert neighborhood

Multistart Local Searches (MLSs): Swap neighborhood (MLS S ) Interchange neighborhood (MLS X ) Insert neighborhood (MLS I ) 4 problems: TSP, LOP, PFSP, QAP 50 instances for each problem of sizes: 10,20,30,40,50,60,70,80,90, repetitions Stopping criterion: 1000n 2 evaluations Experimental design 57

Evaluating the performance of MLSs 58 MLS I MLS X

Correlation Analysis Experiments MLS S MLS X MLS I M ken M cay M ula GM ken GM cay Pearson Correlation Coefficients

Ruggedness of the fitness landscape 60 ProblemSwapInterchangeInsert TSP PFSP LOP QAP The number of local optima for an instance of n=10

278 instances Conclusions The Mallows and Generalized Mallows EDAs under the Kendall’s-τ, Cayley and Ulam distances have despair behaviors in the considered problems. Conducted experiments revealed that there exists a relation between the distances and neighborhoods in EDAs and MLS. The best performing distance-neighborhood is the one that most smooth landscape generates. 61

Studying the linear ordering problem Part II

The linear ordering problem 63

The linear ordering problem 64

The linear ordering problem 65

The linear ordering problem Some applications 66  Aggregation of individual preferences  Kemeny ranking problem  Triangulation of Input-Output tables of the branches of an economy  Ranking in sports tournaments  Optimal weighted ancestry relationships

The insert neighborhood Definitions Two solutions and are neighbors if is obtained by moving an item of from position to position 67

Two solutions and are neighbors if is obtained by moving an item of from position to position The insert neighborhood Definitions 68

Two solutions and are neighbors if is obtained by moving an item of from position to position The insert neighborhood Definitions 69

Two solutions and are neighbors if is obtained by moving an item of from position to position How is the operation translated to the LOP? The insert neighborhood Definitions 70

The linear ordering problem An insert operation 71

The linear ordering problem An insert operation 72

The linear ordering problem An insert operation 73

The linear ordering problem An insert operation 74

The linear ordering problem An insert operation 75

Before After The linear ordering problem An insert operation 76

Before After The linear ordering problem An insert operation 77

Before After Two pairs of entries associated to the item 4 exchanged their position. The linear ordering problem An insert operation 78

Before After The linear ordering problem An insert operation 79 The contribution of the item 4 to the objective function varied from 69 to 61.

The linear ordering problem The contribution of an item to the fitness function 80

The linear ordering problem The contribution of an item to the fitness function 81

The linear ordering problem The contribution of an item to the fitness function 82

Contribution: 54 Vector of differences The linear ordering problem The contribution of an item to the fitness function

Vector of differences Contribution: 89 The linear ordering problem The contribution of an item to the fitness function

The vector of differences Local optima What happens in local optimal solutions? There is no movement that improves the contribution of any item 7 > 0 0 < > > 0 All the partial sums of differences to the left must be positive Depends on the overall solution 85 All the partial sums of differences to the right must be negative

But, Negative sumsPositive sums In order to produce local optima, item 5 must be placed in the first position The vector of differences Local optima 86

The restrictions matrix We propose an algorithm to calculate the restricted positions of the items: 1. Vector of differences. 2. Sort differences 3. Study the most favorable ordering of differences in each positions 87 All the partial sums of differences to the right must be negative

The restrictions matrix We propose an algorithm to calculate the restricted positions of the items: 1. Vector of differences. 2. Sort differences 3. Study the most favorable ordering of differences in each positions 88 All the partial sums of differences to the right must be negative

The restrictions matrix We propose an algorithm to calculate the restricted positions of the items: 1. Vector of differences. 2. Sort differences 3. Study the most favorable ordering of differences in each positions Non-local optima Possible local optima 89

The restrictions matrix 90 Time complexity:

The restricted insert neighborhood Incorporate the restrictions matrix to the insert neighborhood. Discard the insert operations that move items to the restricted positions. Theorem The insert operation that most improves the solution is never restricted. 91

The restricted insert neighborhood 92 Insert neighborhood Restricted Insert neighborhood

The restricted insert neighborhood 93 Insert neighborhood Restricted Insert neighborhood Evaluations: 10 5

The restricted insert neighborhood 94 Insert neighborhood Restricted Insert neighborhood Evaluations: 10 5

The restricted insert neighborhood 95 Insert neighborhood Restricted Insert neighborhood Evaluations: 10 5

The restricted insert neighborhood 96 Insert neighborhood Restricted Insert neighborhood Evaluations: 20 11

The restricted insert neighborhood 97 Insert neighborhood Restricted Insert neighborhood Evaluations: 30 17

The restricted insert neighborhood 98 Insert neighborhood Restricted Insert neighborhood Same final solution Evaluations: 30 17

1000n 2 evals Total MA r vs MA35 (4)31 (8)39 (11)43 (7)41 (9)37 (13)226 (52) ILS r vs ILS37 (2) 49 (1)48 (2)50 (0) 271 (7) 5000n 2 evals Total MA r vs MA37 (2)39 (0)50 (0)49 (1)44 (6) 263 (15) ILS r vs ILS38 (1)36 (3)50 (0)45 (5)46 (4)47 (3)262 (16) 10000n 2 evals Total MA r vs MA39 (0)34 (5)43 (7)50 (0) 49 (1)265 (13) ILS r vs ILS33 (6)37 (2)46 (4)42 (8)43 (7)45 (5)246 (32) 278 instances Experiments Maximum number of evaluations Schiavinotto, T., Stützle, T., The linear ordering problem: instances, search space analysis and algorithms. Journal of Mathematical Modelling and Algorithms instances

Experiments Execution time iterations

278 instances Conclusions A theoretical study of the LOP under the insert neighborhood was carried out. A method to detect the insert operations that do not produce local optima solutions was proposed. As a result, the restricted neighborhood was introduced. Experiments confirmed the validity of the new neighborhood outperforming the two state-of-the-art algorithms. 101

A general multi-objectivization scheme based on the elementary landscape decomposition Part III

Multi-objectivization Definitions Single-objective Problem Elementary landscape decomposition 103 Multi-objective Problem  Aggregation: add new functions.  Introduce diversity  Decomposition: decompose into subfunctions  Optimize separately the subfunctions.

Elementary landscapes Definitions Groover’s wave equation 104 A landscape is An elementary landscape fulfills

Elementary landscape decomposition Conditions If the neighborhood N is SymmetricRegular then the landscape can be decomposed as a sum of elementary landscapes 105 According to Chicano et al. 2010

Elementary Landscape Decomposition The quadratic assignment problem (QAP) Elementary landscape decomposition The quadratic assignment problem (QAP)

Elementary Landscape Decomposition The quadratic assignment problem (QAP) Elementary landscape decomposition The quadratic assignment problem (QAP)

Elementary landscape decomposition 2-objective QAP Generalized QAP 108 QAP According to Chicano et al. 2010

Elementary landscape decomposition 2-objective QAP Generalized QAP 109 According to Chicano et al Landscape 1Landscape 2 Landscape 3 Under the interchange neighborhood

2-objective QAP Elementary landscape decomposition 2-objective QAP Landscape 1 Landscape 2 Landscape In the classic QAP the matrix is symmetric, as a result

Experiments Adapted NSGA-II for the 2-objective QAP SGA for the classical QAP InstancesNSGA-IISGA Random Real-life like73703 Total instances: 35 random, 73 real-life like 111 %68 %95

A general multi-objectivization strategy based on the elementary landscape decomposition was proposed. Based on the decomposition of the QAP under the interchange neighborhood, we reformulated it as a 2-objective problem. Results confirmed that solving the 2-objective QAP formulation is preferred. Specially interesting for the real-life like instances. Conclusions 112

Conclusions and Future Work

Conclusions A new set of EDAs that codify probability models on the domain of permutations has been introduced. –K-order marginals-based models. –The Plackett-Luce model –The Mallows and Generalized Mallows models. Kendall Cayley Ulam The linear ordering problem has been studied and an efficient insert neighborhood system that outperforms existing approaches has been proposed. A general multi-objectivization strategy based on the elementary landscape decomposition has been proposed and applied to solve the quadratic assignment problem. 114

Future Work Part I Investigate mixtures or kernels of Generalized Mallows models to approach multimodal spaces. Study the convergence of the Mallows and Generalized Mallows EDAs to local optima of the implemented distances. Analyze the suitability of the proposed models to solve a given problem by calculating the Kullback-Leibler divergence with respect to the Boltzmann distribution associated to the problem. Include other distances such as Hamming or Spearman. 115

Future Work Part II Investigate multivariate information associated to the items. Study further applications of the restrictions matrix. –Branch and bound algorithms. 116

Future Work Part III Extend the elementary landscape decomposition to the LOP and TSP. –Particular cases of the Generalized QAP. Find an orthogonal basis of functions to decompose the landscape produced by the insert neighborhood under the LOP. Study the shape of elementary landscapes of the decomposition in relation to the values of the QAP instances. 117

Publications Articles J. Ceberio, E. Irurozki, A. Mendiburu & J.A. Lozano (2012). A review on Estimation of Distribution Algorithms in Permutation-based Combinatorial Optimization Problems. Progress in Artificial Intelligence. Vol 1, No. 1, Pp Citations in Google scholar : 30. J. Ceberio, E. Irurozki, A. Mendiburu & J.A. Lozano (2014). A Distance- based Ranking Model Estimation of Distribution Algorithm for the Flowshop Scheduling Problem. IEEE Transactions on Evolutionary Computation. Vol 18, No. 2, Pp J. Ceberio, A. Mendiburu, J.A. Lozano (2015). The Linear Ordering Problem Revisited. European Journal of Operational Research. Vol 241, No. 3, Pp

Publications Articles J. Ceberio, E. Irurozki, A. Mendiburu & J.A. Lozano (2014). A Review of Distances for the Mallows and Generalized Mallows Estimation of Distribution Algorithms. Journal of Computational Optimization and Applications. Submitted. J. Ceberio, A. Mendiburu & J.A. Lozano (2014). Multi-objectivizing the Quadratic Assignment Problem by means of a Elementary Landscape Decomposition. Natural Computing. Submitted. 119

Publications Conference Communications J. Ceberio, A. Mendiburu & J.A. Lozano (2011). A Preliminary Study on EDAs for Permutation Problems Based on Marginal-based Models. In Proceedings of the 2011 Genetic and Evolutionary Computation Conference, Dublin, Ireland, July. J. Ceberio, A. Mendiburu & J.A. Lozano (2011). Introducing the Mallows Model on Estimation of Distribution Algorithms. In Proceedings of the 2011 International Conference on Neural Information Processing, Shanghai, China, November. Pp J. Ceberio, A. Mendiburu & J.A. Lozano (2013). The Plackett-Luce Ranking Model on Permutation-based Optimization Problems.. In Proceedings of the 2013 IEEE Congress on Evolutionary Computation, Cancun, Mexico, June. J. Ceberio, L. Hernando, A. Mendiburu & J.A. Lozano (2013). Understanding Instance Complexity in the Linear Ordering Problem. In Proceedings of the 2013 International Conference on Intelligent Data Engineering and Automated Learning, Hefei, Anhui, China, October. J. Ceberio, E. Irurozki, A. Mendiburu & J.A. Lozano (2014). Extending Distance- based Ranking Models in Estimation of Distribution Algorithms. In Proceedings of the 2014 IEEE Congress on Evolutionary Computation, Beijing, China, 6-11 July. 120

Publications Collaborations E. Irurozki, J. Ceberio, B. Calvo & J. A. Lozano. (2014). Mallows model under the Ulam distance: a feasible combinatorial approach. Neural Information Processing Systems (NIPS) – Workshop of Analysis of Rank Data. 121

Solving Permutation Problems with Estimation of Distribution Algorithms and Extensions Thereof Josu Ceberio