Multi-Agent Systems : 20th European Conference, EUMAS 2023, Naples, Italy, September 14–15, 2023, Proceedings

This volume LNCS 14282 constitutes the refereed proceedings of the 20th European Conference EUMAS 2023, held in Naples, Italy, during September 2023. This volume includes 24 full papers and 5 short papers, carefully selected from 47 submissions. Additionally, the volume features 16 short papers, rigorously reviewed from 20 submissions for the PhD day. The conference focused on the theory and practice of autonomous agents and multi-agent systems, covering a wide range of topics. Preface Organization Invited Speakers Abstracts Game-Theoretic Perspectives in Reactive Synthesis Engineering Social Capabilities in Human-Centered AI Contents Multiple Attribute List Aggregation and an Application to Democratic Playlist Editing 1 Introduction 1.1 Paper Structure 2 Related Work 2.1 Multi-attribute Social Choice 2.2 Proportional Ranking 2.3 Multi-attribute Scheduling 3 Multi-Attribute List Aggregation (MALA) 4 Committee Scoring Rules Using MALA 5 Computational Analysis 6 Algorithms 7 The Democratic Playlist Editing Problem 8 Experimental Analysis 8.1 Experimental Design 8.2 Results and Analysis 9 Outlook References On the Graph Theory of Majority Illusions 1 Introduction 2 Preliminaries 3 Illusions in Arbitrary Networks 4 Illusions in Specific Network Classes 4.1 Graphs with Odd Degrees 4.2 2-Colorable Graphs 4.3 Regular Graphs 5 Conclusion and Outlook References Qualitative Uncertainty Reasoning in AgentSpeak 1 Introduction 2 Background and Related Work 2.1 AgentSpeak 2.2 Uncertainty Reasoning in AgentSpeak 2.3 Dynamic Epistemic Logic 3 Methodology Overview 4 Model Creation: Initializing an Epistemic Model 4.1 Propositionalizing Standard Beliefs 4.2 Propositionalizing Ranged Beliefs 4.3 Propositionalizing Constraints 4.4 Creating the Initial Epistemic Model 5 Model Updates: Updating the Epistemic Model 5.1 Default Belief Event Models 5.2 ``On'' Plans: Overriding Default Event Models 5.3 Creating a DEL Event Model 6 Model Queries: Querying the Epistemic Model 6.1 Evaluating Formulae 7 Operational Semantics 7.1 Model Update Semantics 7.2 Model Query Semantics 8 Application and Evaluation 8.1 Application 8.2 Evaluation 9 Conclusion References JaKtA: BDI Agent-Oriented Programming in Pure Kotlin 1 Introduction 2 Background 2.1 DSL Engineering 2.2 BDI Paradigm and Programming Languages 3 A Kotlin DSL for BDI Agents 3.1 Architecture and Implementation Details 3.2 JaKtA's syntax 4 JaKtA in practice: running example 5 Conclusion, Limitations, and Future Work References Integrating Ontologies and Cognitive Conversational Agents in On2Conv 1 Introduction 2 Background and Related Work 3 On2Conv Design and Implementation 4 Ontology Development and Experiments 5 Conclusions and Future Work References Exploiting Reward Machines with Deep Reinforcement Learning in Continuous Action Domains 1 Introduction 2 Preliminaries 2.1 Reward Machines in RL 2.2 Deep RL Algorithms 3 Adapting Deep RL Algorithms with Reward Machines 3.1 Soft Actor-Critic (SAC) with CRM 3.2 Twin Delayed Deep Deterministic Policy Gradient (TD3) with CRM 4 Experimental Evaluation 4.1 Results in the Half-Cheetah Domain 4.2 Results in the Ant Domain 5 Discussion 6 Conclusion References A Comprehensive Presentation of the Jadescript Agent-Oriented Programming Language 1 Introduction 2 Ontologies in Jadescript 3 Jadescript Agents and Their Behaviours 4 Events and Messages in Jadescript 5 Conclusion References Verifying Programs by Bounded Tree-Width Behavior Graphs 1 Introduction 2 Programs with Recursive Procedure Calls 3 Graphs Representing Program Executions 4 Bounded Tree-Width Analysis of Programs 5 Getting Tree Decompositions for Program Nested Words 6 Implementation 7 Discussion References Behavioral QLTL 1 Introduction 2 Quantified Linear Temporal Logic 3 Skolem Functions for QLTL Semantics 4 Behavioral QLTL (QLTLB) 5 Capturing Advanced Forms of Planning in QLTLB 6 QLTLB Properties 7 QLTLB Satisfiability 8 Conclusion References Lorenzen-Style Strategies as Proof-Search Strategies 1 Introduction 2 Dialogical Logic 2.1 Notation and Terminology 2.2 Dialogical Games 3 Sequent Calculus 3.1 Games on Hyland-Ong Arenas 4 From Dialogical Logic Strategies to Derivations 5 From Derivations to Dialogical Logic Strategies 6 Conclusion and Future Work References SHAPE: A Framework for Evaluating the Ethicality of Influence 1 Introduction 2 Concerns 2.1 Secrecy 2.2 Harm 2.3 Agency 2.4 Privacy 2.5 Exogeneity 3 Governance of Influence 3.1 Professional Culture 3.2 Institutional Ethics Reviews 3.3 Regulation 4 Conclusion References Modelling Group Performance in Multiagent Systems: Introducing the CollabQuest Simulation Game 1 Introduction 2 Group Work as a Mixed-Motive Game 3 Formal Model 3.1 Multiagent Reinforcement Learning 3.2 Task Type and Transition Function 3.3 Payoff System and Expected Reward 4 CollabQuest: Unleashing Collective Potential 5 Conclusions and Future Work References Towards Developing an Agent-Based Model of Price Competition in the European Pharmaceutical Parallel Trade Market 1 Introduction 2 Description of the Parallel Trade Market Model 2.1 Agent-Based Model of Pharmaceutical Parallel Trade Market 2.2 Pricing Strategy 2.3 The Agent-Based Model of Price Competition 3 Case Study 4 Experiments and Results 5 Summary and Outlook References Using a BDI Agent to Represent a Human on the Factory Floor of the ARIAC 2023 Industrial Automation Competition 1 Introduction 2 Agile Robotics for Industrial Automation Competition 2.1 Human Operator Agility Challenge 2.2 Safety Distance Calculation 3 Simulation Software Overview 4 The Human Agent 4.1 Initial Beliefs and Initial Goal 4.2 Plans for Movement Control 4.3 Implementing Personalities 4.4 The Environment Class 4.5 Results and Additional Remarks 5 Conclusions and Future Work References Symbolic ltlf Best-Effort Synthesis 1 Introduction 2 Preliminaries 3 Best-Effort Synthesis Under Environment Specifications 4 Symbolic ltlf Best-Effort Synthesis 4.1 Symbolic dfa Games 4.2 Monolithic Approach 4.3 Explicit-Compositional Approach 4.4 Symbolic-Compositional Approach 5 Empirical Evaluations 5.1 Implementation 5.2 Experiment Methodology 5.3 Experimental Results and Analysis 6 Conclusion References Robust Explanations for Human-Neural Multi-agent Systems with Formal Verification 1 Introduction 2 Background 3 Robust Explanations via Verification 3.1 Robust Counterfactual Explanations 3.2 Robust Semi-factual Explanations 4 Experimental Evaluation 4.1 Automated Credit Scoring 4.2 Traffic Sign Recognition 5 Discussion and Conclusions References ltlf Synthesis Under Environment Specifications for Reachability and Safety Properties 1 Introduction 2 Preliminaries 3 Problem Description 4 Building Blocks for the Algorithms 5 Reachability Tasks, No Env Spec 6 Safety Tasks, No Env Spec 7 Reachability and Safety Tasks, No Env Spec 8 Reachability Tasks, Safety Env Specs 9 Safety Tasks, Safety Env Specs 10 Reachability and Safety Tasks, Safety Env Specs 11 Reachability and Safety Tasks and Env Specs 12 Conclusion References Logic-Based Approximations of Preferences 1 Introduction 2 Hypothetical and Actual Information 3 Approximating Preferences 4 Qualitative Probability and Representation 5 Conclusion References A Comparative Analysis of Multi-agent Simulation Platforms for Energy and Mobility Management 1 Introduction 2 MAS for Integrated Energy and Mobility 2.1 Model Description 2.2 Platform Evaluation Scenario 2.3 Expected Features from MAS Platform 3 Multi-level Selection 3.1 First Round of Selection 3.2 Second Round of Selection 4 Mesa vs. Agents.jl 4.1 Implementation - Model Speed and Scalability Evaluation 4.2 Benchmark Results 5 Conclusion References Observational Preorders for Alternating Transition Systems 1 Introduction 2 An Alternating View of Transition Systems 2.1 Definition 2.2 Strategies and Observations 2.3 Enforcing Preorder 3 Alternating Simulations 3.1 The Simulation Game 4 Strategies as Contexts 4.1 Observational Preorders 4.2 Choice Correspondence and the Morris Preorder 5 The Adequacy Theorems 6 Complements 6.1 The Quest for Symmetric Relations 6.2 Taking Labels Seriously 7 Conclusions and Future Work References Synthesising Reward Machines for Cooperative Multi-Agent Reinforcement Learning 1 Introduction 2 Preliminaries 2.1 Multi-agent Environments 2.2 Reward Machines 2.3 Multi-Agent RL with RMs 2.4 Alternating-Time Temporal Logic 3 Synthesising MGRMs 3.1 Synthesising Reward Machines 3.2 Correctness of the Approach 4 Evaluation 5 Related Works 6 Conclusions and Future Work References Adaptive Cognitive Agents: Updating Action Descriptions and Plans 1 Introduction 2 Background and Related Work 3 Framework 4 Implementation 5 Evaluation 6 Discussion and Future Work References Pretty Good Strategies and Where to Find Them 1 Introduction 2 Preliminaries 2.1 ATL: What Agents Can Achieve 2.2 Models 2.3 Strategies 2.4 Outcome Paths 2.5 Semantics of ATL 2.6 Model Checking and Strategy Synthesis 2.7 Partial Strategies and Strategy Dominance 3 Two Notions of Dominance for Iterated Strategy Improvement 3.1 Multi-criterial Domination: Abstract Template 3.2 Outcome- and Uniformity-Dominance 4 Iterated Strategy Synthesis 5 Experimental Evaluation 5.1 First Benchmark: Random Models 5.2 Second Benchmark: Drone Model 5.3 Running the Experiments 5.4 Results 6 Coalitional Strategies 7 Conclusions References A Multi-agent Sudoku Through the Wave Function Collapse 1 Introduction 2 Background 2.1 Game's Overview 2.2 Math Involvement 2.3 Wave Function Collapse 2.4 Video Game Development as MAS 3 Description and Specification 3.1 Environment Description 3.2 Agent Description 4 Game Implementation and Mechanics 4.1 Spatial Layout of the Grid 4.2 Grid Generation 4.3 Level Generation 4.4 User Interaction 5 Results 6 Conclusions and Future Work References AGAMAS: A New Agent-Oriented Traffic Simulation Framework for SUMO 1 Introduction 2 Background 3 Related Work 4 Agent-Based Traffic Simulation in AGAMAS 4.1 Perceptions and Actions of Agents 4.2 AGAMAS' Architecture 5 Experiments 6 Conclusions and Future Work References Coordinating Systems of Digital Twins with Digital Practices 1 Introduction 2 Preliminary Concepts 2.1 Digital Twins and Agents 2.2 Social Practice Theory 3 Digital Practices: Towards a Design Methodology for Agents and Digital Twins 3.1 Digital Practice 3.2 Designing an Ambient Assisted Living for the Elderly 3.3 Some Considerations 4 Conclusions References On Admissible Behaviours for Goal-Oriented Decision-Making of Value-Aware Agents 1 Introduction 2 Related Work 3 Value Aligning Sequences of Decisions 3.1 State-Level Alignment: Value Preferences 3.2 Plan-Level Alignment and Admissibility 4 Example: Equity in Water Distribution 4.1 Legal and Values Considerations for Water Distribution 4.2 Use Case 5 Conclusions and Future Work References Multi-tasking Resource-Constrained Agents Reach Higher Accuracy When Tasks Overlap 1 Introduction 2 Related Work 3 Experimental Framework 3.1 Environment 3.2 Tasks 3.3 Agents 4 Experiment Outline 4.1 Initial Ontology Induction 4.2 Task Assignment 4.3 Interaction 4.4 Resources Release 4.5 Adaptation 5 Experimental Setting 5.1 Hypothesis 5.2 Parameters 5.3 Measures 6 Results and Discussion 6.1 Statistical Analysis 7 Conclusion References Election Manipulation on Social Networks with Abstention 1 Introduction 2 Definitions 3 Modeling Abstention 4 Election Manipulation 5 Conclusion References Supporting Adaptive Multi-Agent Systems with Digital Twins Environments 1 Introduction 2 Background 2.1 The Role of the Environment in Multi-Agent Systems 2.2 Digital Twins 3 Building Digital Twins Environments 4 Future Works 4.1 Towards a Generalised Model for Adaptive MAS 4.2 Exploiting Digital Twins Capabilities for Adaptation References A Step Forward to Widespread BDI AOP: JaKtA 1 AOP and the Mainstream 2 The Missing Pieces 2.1 Learning Curve vs. Ergonomics 2.2 Tooling 2.3 Middleware/Runtime Requirements 2.4 Concurrency Model 3 The JaKtA Approach 4 The Future Directions References A Brief Overview of an Approach Towards Ethical Decision-Making 1 Introduction 2 Related Work 3 Research Framework 4 Discussion 5 Conclusion References On Verifying Unbounded Client-Server Systems 1 Introduction 1.1 Modeling Unbounded Client-Server Systems 2 Restricted -Nets 3 The Monodic Logic LUCS1 4 Conclusion References Capacity ATL: Reasoning About Agent Profiles and Applications to Cybersecurity 1 Introduction 2 Capacity Alternating-Time Temporal Logic 3 Case Study 4 Future Works References Value-Awareness Engineering: Towards Learning Context-Based Value Taxonomies 1 Introduction and Related Work 2 Value Awareness as Value-Alignment and Admissibility 3 Learning Value-Awareness with Value Taxonomies 4 Conclusions References Virtual Environments via Natural Language Agents 1 Introduction and Motivation 2 Background 3 Related Work 4 Design and Implementation 5 Conclusions and Future Work References Reasoning About Smart Parking 1 Introduction 2 An Algorithmic Solution 3 A Formal Solution 4 Conclusions References Towards the Optimization of Speculative PDES Platforms in Shared-Memory Multi-core Machines 1 Introduction 2 Locality Based Load-Sharing and NUMA Awareness 3 Incremental State Saving Exploiting Memory Protection 4 Effective Access to the Committed Global State 5 Conclusions References Decidability Borders of Verification of Communicating Datalog Agents 1 Introduction 2 CDA Model 3 CDA Verification 4 Quest for Decidability 5 Conclusions References ltlf Best-Effort Synthesis for Single and Multiple Goal and Planning Domain Specifications 1 Introduction 2 Preliminaries 3 Best-Effort Synthesis in Nondeterministic Domains 4 Extensions 4.1 Best-Effort Synthesis in Multiple Planning Domains 4.2 Best-Effort Synthesis for Multiple Goal Specifications 5 Conclusions and Future Work References Neurosymbolic Integration of Linear Temporal Logic in Non Symbolic Domains 1 Introduction 2 Problem Formulation 3 Models 3.1 Recurrent Logic Tensor Networks 3.2 Probabilistic Relaxation of DFA: DeepDFA 3.3 DeepDFA with Probabilistic Grounded Symbols 4 Applications 4.1 Exploiting LTL Knowledge in Image Sequence Classification 4.2 Neural DFA Induction from Traces 4.3 Application to Reinforcement Learning: Visual Reward Machines 5 Conclusions References On Theoretical Questions of Machine Learning, Multi-Agent Systems, and Quantum Computing with Their Reciprocal Applications 1 Introduction 2 Agents and Learning 2.1 Reinforcement Learning 2.2 Graph Neural Networks- GNNs 3 Quantum Computing (QC) Impact on MAS and ML 4 Conclusion References Optimal Rescue Sequences in Disastrous Incidents 1 Introduction 2 Related Works 3 Problem Statement and Contribution 4 Methodology 5 Simulation and Early Results 6 Conclusion and Future Work References Efficient Algorithms for LTLf Synthesis 1 Intorduction 2 Compositional ltlf-to-DFA 3 Forward ltlf Synthesis 4 Conclusion References Agent Behavior Composition in Stochastic Settings 1 Introduction 2 Formalization of the Problem 2.1 Stochastic Policy 2.2 Stochastic Constraint-Based Policy 3 Conclusion and Future Works References Author Index