Applications of Evolutionary Computation: 26th European Conference, EvoApplications 2023 Held as Part of EvoStar 2023 Brno, Czech Republic, April 12–14, 2023 Proceedings

Preface
Organization
Contents
Applications of Evolutionary Computation
An Evolutionary Approach for Scheduling a Fleet of Shared Electric Vehicles
	1 Introduction
	2 Problem Description
	3 Evolutionary Algorithm
		3.1 Encoding
		3.2 Initialization
		3.3 Fitness Evaluation
		3.4 Crossover
		3.5 Mutation
		3.6 Optimization Process
		3.7 Surrogate-Assisted Optimization
	4 Experiments
		4.1 Experimental Setup
		4.2 Parameter Setting and Analysis
		4.3 Experimental Results
	5 Summary and Conclusion
	References
The Specialized Threat Evaluation and Weapon Target Assignment Problem: Genetic Algorithm Optimization and ILP Model Solution
	1 Introduction
	2 Related Work
	3 Problem Description
	4 Method
		4.1 Threat Evaluation with Genetic Algorithm
		4.2 Mathematical Model
		4.3 Application of Physical and Tactical Constraints on Jammer Angle Changes
	5 Experimental Study
		5.1 Data Sets
		5.2 Genetic Algorithm Tests
		5.3 Experiments with the Linear Programming Model
	6 Conclusion
	References
Improving the Size and Quality of MAP-Elites Containers via Multiple Emitters and Decoders for Urban Logistics
	1 Introduction and Motivation
	2 Related Work
	3 Methodology
		3.1 MAP-Elites
		3.2 Set up and Initialisation
		3.3 Representation
		3.4 Emitters
		3.5 Performance Metrics
	4 Decoding Strategies
	5 Managing a Collection of Emitters and Decoders
		5.1 Emitter Selection via Multi-Arm Bandits
	6 Discussion
	7 Conclusion
	References
An Evolutionary Hyper-Heuristic for Airport Slot Allocation
	1 Introduction
	2 Related Work
	3 Solution Approach
		3.1 Allocation Algorithm
		3.2 Request Ordering Heuristics
		3.3 Evolutionary Hyper-heuristic Approach
	4 Experimental Results
		4.1 Individual Constructive Heuristic Performance
		4.2 Hyper-Heuristic Parameter Tuning
		4.3 Hyper-Heuristic Search
	5 Conclusions
	References
A Fitness Landscape Analysis Approach for Reinforcement Learning in the Control of the Coupled Inverted Pendulum Task
	1 The Coupled Inverted Pendulums Task
	2 Robotic Control by Deep Reinforcement Learning
	3 Fitness Landscape Analysis
		3.1 Progressive Random Walk
		3.2 Measure of Ruggedness
		3.3 Measure of Neutrality
	4 (1+1) Evolution Strategies for Reinforcement Learning
	5 Results of the Ruggedness and Neutrality
		5.1 Relation Between Control Parameters and FLA
	6 Conclusion
	References
Local Optima Networks for Assisted Seismic History Matching Problems
	1 Introduction
	2 Method
		2.1 Local Optima Networks
		2.2 Reservoir Model and Fitness Computation
	3 Results
	4 Discussion
	5 Conclusion
	References
Extending Boundary Updating Approach for Constrained Multi-objective Optimization Problems
	1 Introduction
	2 Proposed Method
	3 Numerical Study
		3.1 Osyczka and Kundu (OSY)
		3.2 Bin and Korn Test Problem (BNH)
		3.3 Welded Beam Design Problem
		3.4 Cantilevered Beam Design Problem
	4 Conclusion
	References
Multi-agent vs Classic System of an Electricity Mix Production Optimization
	1 Introduction
	2 Methods
		2.1 Previous Works
		2.2 Classic Approach
		2.3 Multi-agent Based Approach
	3 Experiments
		3.1 Dataset
		3.2 Scenarios
	4 Discussions
	5 Conclusion
	References
A Multi-brain Approach for Multiple Tasks in Evolvable Robots
	1 Introduction
	2 Background
	3 Experimental Work
		3.1 Phenotypes and Simulation
		3.2 Central Pattern Generators
		3.3 HyperNEAT and Compositional Pattern Producing Networks
		3.4 Two-Brain Approach
		3.5 Two-Phase Approach
		3.6 Fitness Functions
		3.7 Experiment Parameters
	4 Results
		4.1 Phase 1: Morphology Evolution
		4.2 Phase 2: Task Learning
		4.3 Combining the Tasks
	5 Discussion
		5.1 Future Work
	6 Conclusion
	References
A Quality-Diversity Approach to Evolving a Repertoire of Diverse Behaviour-Trees in Robot Swarms
	1 Introduction
	2 Background
	3 Methodology
		3.1 Environment, Robots and Simulator
		3.2 Behaviour Tree Representation
		3.3 Algorithms
	4 Results
		4.1 Comparison of Objective Fitness
		4.2 Coverage and QD Scores
	5 Conclusions and Further Work
	References
Evolutionary Based Transfer Learning Approach to Improving Classification of Metamorphic Malware
	1 Introduction
	2 Background
	3 Methodology
		3.1 Creation of the Evolved Malware Mutants
		3.2 Data Collection and Processing
		3.3 NLP Language Models
	4 Experimental Settings
	5 Results
		5.1 Can NLP Language Models Be Used in an Evolutionary-Based Transfer Learning Context to Improve the Classification of Metamorphic Malware?
		5.2 Which of These NLP Models Provides the Best Classification Performance for Metamorphic Malware?
	6 Conclusion
	References
Evolving Lightweight Intrusion Detection Systems for RPL-Based Internet of Things
	1 Introduction
	2 Background
		2.1 RPL
		2.2 Evolutionary Computation
	3 Related Work
	4 Evolving Intrusion Detection Algorithms
	5 Experimental Results
		5.1 Simulation Environment
		5.2 Results
	6 Strengths, Limitations, and Future Directions
	7 Conclusion
	References
A New Prediction-Based Algorithm for Dynamic Multi-objective Optimization Problems
	1 Introduction
	2 Dynamic Multi-objective Evolutionary Algorithms
	3 The Proposed Algorithm
		3.1 Change Detection Strategy
		3.2 Change Response Strategy
		3.3 Prediction Strategies for the Individuals
		3.4 Multiple Reactions
	4 Experimental Design
		4.1 Test Problems
		4.2 Performance Metrics
		4.3 Algorithms in the Empirical Study
	5 Experimental Results
	6 Conclusions
	References
Epoch-Based Application of Problem-Aware Operators in a Multiobjective Memetic Algorithm for Portfolio Optimization
	1 Introduction
	2 Materials and Methods
		2.1 The Optimization Problem
		2.2 A Memetic Approach
		2.3 Data Used in the Analysis
	3 Results
		3.1 Experimental Setup
		3.2 Sensitivity Analysis
		3.3 Performance Comparison
	4 Conclusions
	References
Reducing the Price of Stable Cable Stayed Bridges with CMA-ES
	1 Introduction
	2 Related Work
	3 Problem Definition
	4 The Approach
	5 Experimental Results
	6 Conclusions
	References
On the Evolution of Boomerang Uniformity in Cryptographic S-boxes
	1 Introduction
	2 Background
		2.1 Notation
		2.2 S-boxes
	3 Related Work
	4 Experimental Setup
		4.1 Encodings
		4.2 Fitness Functions
		4.3 Algorithms and Parameters
	5 Experimental Results
		5.1 Integer Encoding
		5.2 Permutation Encoding
		5.3 CA-Based Encoding
		5.4 Representation Comparison
		5.5 Multi-objective Optimization
	6 Conclusions and Future Work
	References
Using Knowledge Graphs for Performance Prediction of Modular Optimization Algorithms
	1 Introduction
	2 Formal Representation of Modular Optimization Algorithms
	3 Performance Prediction via KG Triple Classification
		3.1 Construction of the KG
		3.2 KG Embedding-Based Pipeline for Automated Algorithm Performance Prediction
	4 Evaluation Results
		4.1 Leave Random Performance Triplets Out Validation
		4.2 Leave Problem Instances/algorithm Configurations Out Validation
	5 Addressing the Problem of Imbalanced Classification
	6 Conclusions and Future Work
	References
Automatic Design of Telecom Networks with Genetic Algorithms
	1 Introduction
	2 Background
		2.1 Related Work
	3 Proposed Approach
		3.1 Representation
		3.2 Variation Operators
		3.3 Local Search
		3.4 Fitness Function
	4 Experimental Setup
	5 Experimental Results
	6 Conclusion
	References
RF+clust for Leave-One-Problem-Out Performance Prediction
	1 Introduction
	2 Related Work
	3 LOPO Algorithm Performance Prediction
	4 Experimental Design
	5 Results
	6 Conclusions
	References
Evolving Non-cryptographic Hash Functions Using Genetic Programming for High-speed Lookups in Network Security Applications
	1 Introduction
	2 Related Work
	3 Design Criteria
	4 Design and Implementation
		4.1 Fitness Function
		4.2 Terminal and Non-terminal Set
		4.3 Parameters Setting
		4.4 Stopping Criteria
		4.5 Software and Hardware Implementation
	5 Results and Discussion
		5.1 Software Evaluation
		5.2 Hardware Evaluation
	6 Conclusion
	References
Use of a Genetic Algorithm to Evolve the Parameters of an Iterated Function System in Order to Create Adapted Phenotypic Structures
	1 Introduction
	2 Motivation for the Research
	3 Design of the Algorithm
	4 Results
	5 Further Test and Evaluation
	6 Conclusions
	References
Analysis of Evolutionary Computation Methods: Theory, Empirics, and Real-World Applications
To Switch or Not to Switch: Predicting the Benefit of Switching Between Algorithms Based on Trajectory Features
	1 Introduction
	2 Background
		2.1 BBOB
		2.2 ELA
		2.3 DynAS
	3 Experimental Setup
		3.1 Algorithm Portfolio
		3.2 Finding Usecases Using Irace
	4 Predicting Benefits of Switching
		4.1 Setup
		4.2 Results
		4.3 Impact of Features
	5 Conclusions and Future Work
	References
Frequency Fitness Assignment on JSSP: A Critical Review
	1 Replication Studies: Unpopular and Necessary
	2 The Job Shop Scheduling Problem and Its Algorithms
	3 The Original and the Replication
	4 Results
	5 Conclusion and Discussion
	References
A Collection of Robotics Problems for Benchmarking Evolutionary Computation Methods
	1 Introduction
	2 Forward Kinematics of a Robotic Arm
	3 Definition of the Benchmark Problems
	4 Selected Algorithms for the Numerical Investigation
	5 Numerical Investigation
	6 Exploratory Landscape Analysis
	7 Conclusion
	References
BBOB Instance Analysis: Landscape Properties and Algorithm Performance Across Problem Instances
	1 Introduction
	2 Related Work
		2.1 COCO and the BBOB Benchmark Suite
		2.2 Exploratory Landscape Analysis
	3 Experimental Setup and Reproducibility
	4 Instance Similarity Using ELA
	5 Algorithm Performance Across Instances
	6 Properties Across Instances
	7 Conclusions and Future Work
	References
A Fitness-Based Migration Policy for Biased Random-Key Genetic Algorithms
	1 Introduction
	2 Background and Literature Review
		2.1 Biased Random-Key Genetic Algorithms
		2.2 Distributed Genetic Algorithms
		2.3 Literature Review
	3 Proposed Method
	4 Experiments, Results and Analysis
	5 Conclusion and Future Works
	References
Nullifying the Inherent Bias of Non-invariant Exploratory Landscape Analysis Features
	1 Introduction
	2 Exploratory Landscape Analysis
	3 Invariance in Exploratory Landscape Analysis
	4 Black-Box Optimization Benchmark
	5 Transformation Function
	6 BBOB Function Prediction
	7 Automated Algorithm Selection
	8 Conclusions and Outlook
	References
A Robust Statistical Framework for the Analysis of the Performances of Stochastic Optimization Algorithms Using the Principles of Severity
	1 Introduction
	2 Background
		2.1 Statistical Analysis
		2.2 Existing Statistical Tools to Compare Algorithm Performances
	3 Severity
		3.1 Other Existing Alternative Measures
	4 Proposed Framework: AlgCompare
	5 Casestudy
		5.1 CMA-ES Outperforms RSPSO
		5.2 RSPSO Outperforms CMA-ES
		5.3 Evaluating Other Measures
	6 Summary and Outlook
	References
Towards Constructing a Suite of Multi-objective Optimization Problems with Diverse Landscapes
	1 Introduction
	2 Background
		2.1 Multi-objective Optimization Problems
		2.2 The bbob Functions
		2.3 ELA Features
	3 Methodology
		3.1 Generating the Base Set of Problems
		3.2 Computing the ELA Features
		3.3 Selecting a Diverse Subset of Problems
		3.4 Generating Similar Instances
	4 Results and Discussion
		4.1 Proof of Concept on 2-D Problems
		4.2 Results on Multiple Dimensions
		4.3 Limitations
	5 Conclusions
	References
Computational Intelligence for Sustainability
Using Genetic Programming to Learn Behavioral Models of Lithium Batteries
	1 Introduction
	2 ECMs of a Lithium Titanate Oxide Battery
		2.1 Adopted ECM for Lithium Titanate Oxide Battery
		2.2 ECM-Based Simulator
		2.3 Dataset
	3 Behavioral Modeling of LTO Battery
		3.1 GP-Based Modeling
		3.2 NN-Based Modeling
	4 Results and Discussion
	5 Conclusions
	References
An Intelligent Optimised Estimation of the Hydraulic Jump Roller Length
	1 Introduction
	2 Hydraulical Aspects
		2.1 The Hydraulic Jump over a Rough Surface
		2.2 Definition of a Roughness Height Modelling Function
	3 The Learning Scheme
	4 Black-Box Optimisers
		4.1 Random and Quasi-random Search
		4.2 Simple Evolution Strategies
		4.3 Covariance Matrix Adaptation Evolution Strategies
		4.4 Differential Evolution
		4.5 Particle Swarm Optimization
		4.6 Nelder-Mead
	5 Experiments
		5.1 Experimental Setup
		5.2 Experimental Results
	6 Conclusion and Future Work
	References
.26em plus .1em minus .1emA Multispectral Image Classification Framework for Estimating the Operational Risk of Lethal Wilt in Oil Palm Crops
	1 Introduction
	2 Materials and Methods
		2.1 Acquiring and Preparing the Data Set
		2.2 Defining Losses and Operational Risk
		2.3 The Employed Vegetation Indices
		2.4 A Hybrid Neural Deep Learning System
		2.5 Evaluation Metrics
	3 Experimental Setup and Validation
		3.1 Assessment Scenarios
		3.2 Quantifying the Losses
	4 Experimental Results
	5 Conclusions and Future Work
	References
An AI-Based Support System for Microgrids Energy Management
	1 Introduction
		1.1 Aim and Objectives
	2 Materials and Methods
		2.1 Data
		2.2 Simulation of Micro-Grid for Validation
		2.3 Machine Learning Algorithms
		2.4 Evaluation Metrics
	3 Results and Discussion
	4 Conclusions and Future Work
	References
Predicting Normal and Anomalous Urban Traffic with Vectorial Genetic Programming and Transfer Learning
	1 Introduction
	2 Background
	3 Motivating Scenario
		3.1 Traffic Anomaly Handling over Time: Skipping and Smoothing
		3.2 Traffic Anomaly Handling Across Space: Transfer Learning
	4 GENTLER Explained
	5 Experimental Investigation
		5.1 Sensor Selection and Traffic Readings Analysis
		5.2 Results Discussion
	6 Conclusions
	References
Evolutionary Computation in Edge, Fog, and Cloud Computing
Energy-Aware Dynamic Resource Allocation in Container-Based Clouds via Cooperative Coevolution Genetic Programming
	1 Introduction
	2 Related Work
	3 Problem Formulation
	4 A Hybrid Approach for Dynamic RAC
		4.1 Representation, Terminal and Function Set
		4.2 Fitness Evaluation
	5 Experiment
		5.1 New Dataset
		5.2 Parameter Settings
		5.3 Performance Comparison
		5.4 Further Analysis
	6 Conclusions
	References
A Memetic Genetic Algorithm for Optimal IoT Workflow Scheduling
	1 Introduction
	2 Related Works
	3 Problem Formulation
		3.1 Makespan Model
		3.2 Energy Consumption Model
	4 Algorithm Design
		4.1 Solution Representation and Population Initialization
		4.2 Fitness Evaluation
		4.3 Evolution Operators
		4.4 Local Search
	5 Experiment Results
		5.1 Benchmark Problems
		5.2 Parameter Settings
		5.3 Results
	6 Conclusions
	References
Multi-objective Location-Aware Service Brokering in Multi-cloud - A GPHH Approach with Transfer Learning
	1 Introduction
	2 Related Works
	3 Problem Definition
	4 GPHH with Transfer Learning for MOLSB
		4.1 Representation and Terminal Set
		4.2 Transfer Approaches
		4.3 Crossover and Mutation
		4.4 Fitness Evaluation
	5 Experiments
		5.1 Simulation and Datasets
		5.2 Baseline Algorithm
		5.3 Parameter Settings
		5.4 Results
	6 Conclusion
	References
Evolutionary Machine Learning
Evolving Neural Networks for Robotic Arm Control
	1 Introduction
	2 Related Work
		2.1 Neuroevolution
		2.2 Robotic Arm Controllers
	3 Experimental Setup
		3.1 Simulated Robotic Arm
		3.2 Experiments
		3.3 Evolving Robotic Arm Controller
		3.4 Hyperparameters
	4 Results
		4.1 Comparing CMA-ES, DE and PSO for Single-Step Task
		4.2 Comparing Single-Step with Mult-Step Control
		4.3 Learning from Expert Demonstrations
		4.4 Movement with Disabled Joints
		4.5 Pick and Place Task Using Supervisor Network
	5 Conclusion and Future Work
	References
Centroid-Based Differential Evolution with Composite Trial Vector Generation Strategies for Neural Network Training
	1 Introduction
	2 Background
		2.1 Feedforward Neural Networks
		2.2 Composite Differential Evolution
		2.3 Centroid-Based Sampling Strategy
	3 Proposed Cen-CoDE Algorithm
		3.1 Representation
		3.2 Objective Function
		3.3 Centroid-Based CoDE
		3.4 Algorithm
	4 Experimental Results
	5 Conclusions
	References
AMTEA-Based Multi-task Optimisation for Multi-objective Feature Selection in Classification
	1 Introduction
	2 Related Work
		2.1 Evolutionary Multi-objective Feature Selection
		2.2 Multi-task Optimisation for Feature Selection
		2.3 Adaptive Model-Based Transfer-Enabled Evolutionary Algorithm–AMTEA
	3 Proposed Method
		3.1 The Framework of FSMTO
		3.2 Build an Online Knowledge Pool
		3.3 Transform the Probabilistic Models
	4 Experiment Design
		4.1 Benchmark Datasets
		4.2 Parameter Settings
		4.3 Result Analyses and Discussions
	5 Conclusions
	References
Under the Hood of Transfer Learning for Deep Neuroevolution
	1 Introduction
	2 Background Knowledge
		2.1 Grammar-based Neuroevolution
		2.2 Incremental Development/Transfer Learning
		2.3 Neuroevolution Trajectory Networks
	3 Experimental Methodology
	4 Performance Analysis
	5 NTNs Analysis
		5.1 Sampling and Model Construction
		5.2 NTNs Analysis: Revealing Transfer Learning
	6 Conclusion
	References
Feature Selection on Epistatic Problems Using Genetic Algorithms with Nested Classifiers
	1 Introduction
	2 Methodology
		2.1 Datasets
		2.2 Defining Baselines
		2.3 Classifiers
		2.4 Adding Feedback
	3 Experimental Setup
	4 Results
		4.1 10 Feature Datasets
		4.2 100 Feature Datasets
		4.3 1000 Feature Datasets
		4.4 Effect of the Feedback Mechanism: GAMDT vs GAMDT+
	5 Conclusions
	References
Grammar-Guided Evolution of the U-Net
	1 Introduction
	2 Background
		2.1 Image Segmentation
		2.2 Convolutional Neural Networks
		2.3 Autoencoders
		2.4 The U-Net and SA-UNet
		2.5 Evolutionary Algorithms
		2.6 Grammatical Evolution
	3 Experimental Setup
		3.1 Dataset
		3.2 Hardware and Software
		3.3 Evaluation Metrics
		3.4 Evolutionary Process
	4 Results
	5 Conclusion
	References
Explaining Recommender Systems by Evolutionary Interests Mix Modeling
	1 Introduction
	2 Recommender Systems
	3 Understanding the Recommendations
		3.1 Simple Regression Model
		3.2 Saturation Regression Model
		3.3 Interpreting the Saturation Regression Model
		3.4 Problem Definition
	4 Evolutionary Interests Mix Modeling Algorithm (EIMMA)
		4.1 Search Space and Objective Functions
		4.2 Crossover and Mutation
		4.3 Selection
		4.4 EIMMA
	5 Experiments
		5.1 Dataset
		5.2 Case Study
		5.3 Summary of Results
	6 Conclusions
	References
Machine Learning and AI in Digital Healthcare and Personalized Medicine
Multi-objective Evolutionary Discretization of Gene Expression Profiles: Application to COVID-19 Severity Prediction
	1 Introduction
	2 Background
		2.1 Feature Selection
		2.2 Gene Expression Profiles
	3 Proposed Approach
		3.1 Feature Selection
		3.2 Multi-objective Evolutionary Discretization
	4 Experimental Evaluation
		4.1 Data
		4.2 Feature Selection
		4.3 Profile Generator
	5 Discussion
	6 Conclusions and Future Works
	References
Interactive Stage-Wise Optimisation of Personalised Medicine Supply Chains
	1 Introduction
	2 Supply Chain and Problem Formulation
	3 Related Work
		3.1 Problem Stages
	4 Solution Methods
	5 Case Study
	6 Results
	7 Conclusion and Future Research Directions
	References
Resilient Bio-inspired Algorithms
Further Investigations on the Characteristics of Neural Network Based Opinion Selection Mechanisms for Robotics Swarms
	1 Introduction
	2 Methods
	3 Results
		3.1 Test I: The Symmetry-Breaking Test
		3.2 Test II: The Tiles Distribution Test
		3.3 Test III: Robots' Floor Sensor Readings Failure
	4 Conclusion
	References
Soft Computing Applied to Games
Deep Reinforcement Learning for 5 5 Multiplayer Go
	1 Introduction
	2 Multiplayer Go
	3 Deep Reinforcement Learning
		3.1 Monte Carlo Tree Search
		3.2 AlphaZero
		3.3 Descent
	4 Experimental Results
		4.1 Training of AlphaZero and Descent
		4.2 Black Against White and Red
	5 Conclusion
	References
Genetic Programming and Coevolution to Play the Bomberman™ Video Game
	1 Introduction
	2 Related Work
	3 Game Environment
	4 GP Setup
		4.1 Building Blocks
		4.2 Coevolutionary Setup
		4.3 Fitness Vector Comparison
		4.4 Parameters
	5 Results
	6 Conclusions
	References
Surrogate-Assisted Evolutionary Optimisation
A Surrogate Function in Cellular GA for the Traffic Light Scheduling Problem
	1 Introduction
	2 Related Work
	3 Algorithmic Proposal
		3.1 Problem Definition
		3.2 Characterizing Cellular Genetic Algorithms
		3.3 Representation and Solution Strategy
		3.4 Surrogate Model
	4 Experiments and Results
		4.1 Datasets, Parameter Settings, and Environment
		4.2 Results and Analysis
	5 Conclusions
	References
Surrogate-Assisted (1+1)-CMA-ES with Switching Mechanism of Utility Functions
	1 Introduction
	2 Backgrounds
		2.1 CMA-ES with Surrogate Models
		2.2 Utility Transformations of Objective Function Value
	3 Surrogate-Assisted CMA-ES with Elitist Strategy
		3.1 Gaussian Process CMA-ES
		3.2 Warped Surrogate CMA-ES
	4 Switching Gaussian Process CMA-ES
		4.1 Utility Functions
		4.2 Switching Mechanism of Utility Function
		4.3 Implementation of SGP-CMA-ES
	5 Experiment
		5.1 Experimental Setting
		5.2 Results on Parametrized Benchmark Functions
		5.3 Results with Transformation Using Power Function
		5.4 Results with Transformation Using Floor Function
	6 Conclusion
	References
Author Index