Optimization and Learning: 5th International Conference, OLA 2022, Syracuse, Sicilia, Italy, July 18–20, 2022, Proceedings

Preface
Organization
Contents
Optimization and Learning
Evolutionary-Based Co-optimization of DNN and Hardware Configurations on Edge GPU
	1 Introduction and Related Works
	2 Motivation
	3 Proposed Approach
		3.1 Problem Formulation
		3.2 Optimization Methodology
	4 Evaluation
		4.1 Experimental Setup
		4.2 Experimental Results
	5 Conclusion
	References
Maximum Information Coverage and Monitoring Path Planning with Unmanned Surface Vehicles Using Deep Reinforcement Learning
	1 Introduction
	2 State of the Art
	3 Methodology
		3.1 Sequential Decision Problem
		3.2 DRL Framework
	4 Simulations and Results
		4.1 Evaluation Metrics
		4.2 Learning Results
		4.3 Metric Comparison with Other Methods
	5 Conclusions
	References
Tuning ForestDisc Hyperparameters: A Sensitivity Analysis
	1 Introduction
	2 Related Work
		2.1 ForestDisc Algorithm
		2.2 ForestDisc Tunning Parameters
	3 ForestDisc Sensitivity Analysis
		3.1 Experimental Set up
		3.2 ForestDisc Sensitivity to the Number of Trees and the Non-linear Optimization Algorithm Used
		3.3 ForestDisc Sensitivity to the Number of Moments Used
	4 Conclusion
	References
Multi-objective Hyperparameter Optimization with Performance Uncertainty
	1 Introduction
	2 GPR and TPE: Basics
	3 Proposed Algorithm
	4 Numerical Simulations
	5 Results
	6 Concluding Remarks
	References
Novel Optimization Techniques
A New Algorithm for Bi-objective Problems Based on Gradient Information
	1 Introduction
	2 The MGDA Algorithm
	3 The Proposed Retro-MGDA Algorithm
		3.1 Principle and Motivation
		3.2 Controlling Moves on the Objective Space by Using the Gradient Information
		3.3 Reto-MGDA Structure
	4 Parameters Setting
	5 Experiments and Results
		5.1 Performance Measures
		5.2 Numerical Results
	6 Conclusion
	References
Adaptive Continuous Multi-objective Optimization Using Cooperative Agents
	1 Problem Statement and Positioning
		1.1 Search Space Topology
		1.2 Multi-objective Optimization Approaches
	2 AMAS Theory for Optimization
		2.1 Natural Domain Modeling
		2.2 Agent Internal State and Behavior
	3 Photonics Problem Modeling and Implementation
	4 Experiments
	5 Conclusion and Perspectives
	References
Integer Linear Programming Reformulations for the Linear Ordering Problem
	1 Introduction
	2 Problem Statement and Reference ILP Formulation
	3 From ATSP to Consensus Ranking, Tighter Formulations
	4 Other ILP Reformulations
		4.1 ILP Formulation with O(N2) Variables and Constraints
		4.2 Three-Index Flow Formulation
		4.3 Another Three-Index Flow Formulation
	5 Computational Experiments and Results
		5.1 Data Generation and Characteristics
		5.2 Comparing LP Relaxations
		5.3 Comparing Branch and Bound Convergences
		5.4 Variable Fixing Heuristics
	6 Conclusions and Perspectives
	References
SHAMan: A Versatile Auto-tuning Framework for Costly and Noisy HPC Systems
	1 Introduction
	2 Related Works and Software
	3 Theoretical Background
		3.1 An Overview of the Optimization Loop
		3.2 Stop Criteria
		3.3 Resampling for Noisy Systems
		3.4 Pruning of Expensive Systems
	4 Software Architecture and Features
		4.1 Terminology
		4.2 Optimization and Vizualization Procedure
		4.3 Software Architecture
		4.4 Implementation Choices
	5 Use-Cases and Results
		5.1 I/O Accelerators
		5.2 Tuning MPI Collectives
	6 Conclusion
	References
Cooperation-Based Search of Global Optima
	1 Introduction
	2 Literature Review
	3 CoBOpti: Cooperation-Based Optimization
		3.1 General Principle
		3.2 Local Search
		3.3 Semi-local Search
		3.4 Cooperation Mechanisms
	4 Experiments and Results
		4.1 Test Functions
		4.2 Methods Comparison
		4.3 Results
	5 Analysis and Discussion
	6 Conclusion
	References
Data-Driven Simulation-Optimization (DSO): An Efficient Approach to Optimize Simulation Models with Databases
	1 Introduction
		1.1 Research Motivation
		1.2 Novelty and Main Contributions of the Paper
		1.3 Organization and Structure of the Paper
	2 Literature Review and Background
	3 The Proposed DSO Platform
	4 Experimental Analysis: DSO for Job Scheduling and Sequencing
		4.1 Experiment 1: Job Scheduling with DSO
		4.2 Experiment 2: Job Sequencing with DSO
	5 Conclusion and Future Works
	A Appendix A
	References
Logistics
Sweep Algorithms for the Vehicle Routing Problem with Time Windows
	1 Introduction
	2 Mathematical Formulation
	3 Sweep Strategy
		3.1 Traditional Sweep
		3.2 Sweep Algorithm Depending on Time Window Length
		3.3 Sweep Algorithm Depending on Time Window Length and Overall Capacity
	4 Computational Results
	5 Conclusion
	References
Improving the Accuracy of Vehicle Routing Problem Approximation Using the Formula for the Average Distance Between a Point and a Rectangular Area
	1 Introduction
	2 Literature Review
	3 Approximation of the Average Distance Between a Point and an Area
		3.1 Problem Setting
		3.2 Model
	4 Experimental Setting and Results
		4.1 Experimental Setting
		4.2 Results
	5 Conclusions
	References
Optimal Delivery Area Assignment for the Capital Vehicle Routing Problem Based on a Maximum Likelihood Approach
	1 Introduction
	2 Formulation
		2.1 Our Concept
		2.2 Formulation Based on Demand Points
		2.3 Formulation Based on Zones
	3 Data Preparation
	4 Results
	5 Discussion
	6 Conclusion
	References
Neural Order-First Split-Second Algorithm for the Capacitated Vehicle Routing Problem
	1 Introduction
	2 Problem Statement
	3 Related Work
		3.1 Neural Combinatorial Optimization (NCO) for the CVRP
		3.2 Two-Step Algorithms for the Vehicle Routing Problem
		3.3 Graph Neural Networks
	4 The Neural Order-First Split-Second Algorithm
		4.1 Instance Features
		4.2 NOFSS Encoding-Decoding Architectures
	5 Experiments
		5.1 Comparison with a Full-Learning Setting
		5.2 Comparison to Handcrafted Heuristics
		5.3 Influence of the Type of Encoder
		5.4 On Models Generalization
	6 Conclusion
	References
Applications
GRASP-Based Hybrid Search to Solve the Multi-objective Requirements Selection Problem
	1 Introduction
	2 Requirements Selection
		2.1 Related Work
		2.2 Multi-objective Formulation
	3 Proposal
		3.1 GPPR: A GRASP Algorithm with Pareto Front and Path Relinking
		3.2 Solution Encoding
		3.3 Construction
		3.4 Local Search
		3.5 Path Relinking
	4 Evaluation Setup
		4.1 Algorithms
		4.2 Datasets
		4.3 Methodology
	5 Results and Analysis
		5.1 Best Configurations
		5.2 Pareto Results
		5.3 Metrics Results
	6 Conclusions and Future Work
	References
Comparing Parallel Surrogate-Based and Surrogate-Free Multi-objective Optimization of COVID-19 Vaccines Allocation
	1 Introduction
	2 COVID-19 Vaccine Distribution Problem
	3 Parallel Multi-objective Evolutionary Algorithms
		3.1 Variation Operators of Evolutionary Algorithms
		3.2 Parallel Multi-objective Evolutionary Algorithms
	4 Experiments
	5 Conclusion
	References
Decentralizing and Optimizing Nation-Wide Employee Allocation While Simultaneously Maximizing Employee Satisfaction
	1 Introduction
	2 State of the Art
	3 Methodology
		3.1 Problem Formulation
		3.2 Problem Representation
		3.3 Overview of the Objectives
		3.4 Modelling Satisfaction Function
		3.5 Modelling Dispersion Function
	4 Experiments and Results
		4.1 Results of Satisfaction Modelling
		4.2 Results of Optimization
	5 Conclusion
	References
Categorical-Continuous Bayesian Optimization Applied to Chemical Reactions
	1 Introduction
	2 Problem Definition
	3 Propositions
		3.1 Gaussian Process Kernel
		3.2 Acquisition Function Optimization
	4 Results
		4.1 Acquisition Function Optimizer
		4.2 Comparison with Other Methods
		4.3 Kernel Influence on Performances
	5 Conclusion
	References
Assessing Similarity-Based Grammar-Guided Genetic Programming Approaches for Program Synthesis
	1 Introduction
	2 Background and Related Work
		2.1 Genetic Programming
		2.2 Grammar-Guided Genetic Programming
		2.3 Problem Text Description To/From Source Code
	3 Similarity-Based G3P
		3.1 Proposed Approach
		3.2 Program Similarity Assessment Approaches
	4 Experiment Setup
		4.1 General Program Synthesis Benchmark Suite
		4.2 Target Programs
		4.3 G3P Parameter Settings
	5 Results
		5.1 Comparison of Similarity Measures
		5.2 Comparison Against Error Rate-Based G3P
	6 Conclusion and Future Work
	References
Author Index