Advances in Real-Time Artificial Intelligence and Multi-Agent Systems

Dear Colleagues, 

PREMISES

The study of task scheduling in real time has brought about a Copernican revolution in the conception of the correctness of an algorithm. Under a more holistic perspective, the evaluation of the adequacy of a computational process is not limited to the verification of the correctness of the implemented algorithm but also takes into account the response time of the actual process by evaluating its appropriateness according to the its goals. In other words, in addition to the requirement of logical correctness, which is desirable for any program, real-time correctness must be evaluated according to the respect of temporal constraints, precedence constraints and constraints forced by the sharing of mutually exclusive resources with other interacting processes. Therefore, this holistic perspective cannot disregard the analysis of the operational context in which the single process operates, both software and hardware. In hard-real time application contexts where the failure of an action or the simple missing of a deadline can have catastrophic consequences, the behavioral adequacy of a control system cannot be established by application testing alone.

REAL-TIME ARTIFICIAL INTELLIGENCE

Achieving intelligent behavior in these contexts has proven problematic mainly because classical Artificial Intelligence (AI) algorithms, such as logical reasoning and search, completely disregard the measurement of time passing while searching for a solution, and this is even more serious when reasoning and action must be alternated. Since when AI applications became mature, there has been growing interest in applying them into complex systems and physical equipment which involve hard deadlines, especially in Cyber-Physical Systems (CPS) scenarios. Unfortunately, most AI algorithms are characterized by unpredictable or high-variance performances, making them unsuitable for real-time control under soft and/or hard deadlines. In other words, the real-time response requirement has been considered mostly incompatible with the unpredictability of classical AI, namely its inability to accurately time its performance.

However, recently research has been conducted on tailoring AI techniques to make them more predictable by explicitly reasoning “within and about” strict timing and precedence constraints between different tasks.

So, a first aim of this special issue is to offer an opportunity to researchers who are using classical AI algorithms in practical application contexts, where time constraints need to be guaranteed.

REAL-TIME MULTI-AGENT SYSTEMS

However, little effort has been spent to transfer these approaches over Multi-Agents Systems (MAS) where additional constraints deriving from concurrent use of mutually exclusive resources stand (e.g. internal memory, communication channels and peripherals such as sensors and actuators).

MAS have been a relevant topic within AI since its very beginning, and their technological advancements lead to a concrete adoption of decentralized flexible systems with increasing connections, interactions, and computational capabilities. Today radically new challenges are arising from the domains of the “Internet of Things” (IoT), CPS and “safety-critical” systems. Unfortunately, in these regards MAS tend to reproduce the same myopic approach of their parent discipline: high-quality of reasoning and human-like interaction with little attention to concrete temporal and resources constraints. In “safety-critical” systems, MAS should not only exhibit rational human-standard behaviors, they must also guarantee the completions of tasks within their deadlines without violating priorities and precedences constraints in accessing mutually exclusive resources. Furthermore, since agents interact, negotiate resources and exchange executions of tasks in a social manner, real-time guarantees should not only be provided at the level of each single agent but they should also be evaluated at the level of the "emergent behaviour" of the entire Agency, and this last vision is quite challenging from a theoretical point of view.

So, a second aim of this special issue is that of gathering contributions from both theoretical and pragmatic perspectives, targeting the employment of MAS in IoT and CPS through the exploitation of methodologies, algorithms and applications from the Real Time Community. We are looking at the next-generation Intelligent CPS, capable to face the challenges of the “ever more connected” IoT era.

Topics of interest include, but are not limited to, the following:

• Real-Time Problem Solving·       
• Real-Time Artificial Intelligence Control·        
• Real-Time Multi-Agent Systems·        
• Real-Time Autonomous Systems·        
• Real-Time Distributed Problem Solving·        
• Real-Time Smart Task Allocation and Execution·        
• Real-Time Negotiation and Interaction Protocols·        
• Simulators and Architectures for Real-Time MAS·        
• Agent-Oriented Programming for Control Applications·        
• Agent-Oriented Programming for the Internet Of Things·        
• Agent-Oriented Programming for Cyber-Physical Systems·        
• Agent Oriented Software Engineering for Real-Time Systems·        
• Cyber-Physical Agents·        
• Emergent Behavioral analysis·        
• Real-Time Behavior Scheduling·        
• Robot and Multi-robot Systems·        
• Simulation of Multi-Agent Systems·        
• Real-Time Cooperation and Coordination·        
• Agent Platforms for safety-critical systems·        
• Performance analysis of Multi-Agent Systems·        
• Response Time analysis in Multi-Agent Systems·       
• Industrial applications of Multi-Agent Systems·        
• Intelligent Real-Time Control for Power Distribution·        
• Intelligent Real-Time Control for Healthcare·        
• Control in Real-Time Bidding·        
• Edge Computing and Artificial Intelligence 

Technical Program Committee Members：

1. Prof. Paulo Leitao Department of Electrical Engineering,Campus Santa Apolónia, Apartado 1134
5301-857 Bragança,Portugal

email: pleitao@ipb.pt

2. Prof. Tommaso Cucinotta Institute of Communication, lnformation and Perception Technologies  

email: tommaso.cucinotta@santannapisa.it

3. Dr. Davide Calvaresi Institute of Information Systems

email: davide.calvaresi@hevs.ch

4. Prof. Carlos Carrascosa Casamayor School of Informatics,Universitat Politècnica de València,
Camino de Vera, s/n, 46022 Valencia, Valencia, Spain

email: carrasco@dsic.upv.es

Prof. Dr. Aldo Franco Dragoni
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.