[agents] CFP: MassiveMAS at AAMAS 2015 (Autonomous Agents and Multi-Agent Systems at Scale)

[Long Tran-Thanh]

[Apologies for multiple copies of this announcement]

------------------------------------------------------------------------
Preliminary Call for papers
MassiveMAS 2015
First International Workshop on
Autonomous Agents and Multi-Agent Systems at Scale
https://massivemas.wordpress.com/
------------------------------------------------------------------------
Held in conjunction with AAMAS 2015
May 4-8, 2015, Istanbul, Turkey
The MassiveMAS workshop will take place on May 4 or 5
------------------------------------------------------------------------

*Workshop description

Many recent real-world problems could be modelled as a massive 
multi-agent systems. Examples range from smart grids and crowdsourcing 
systems, to financial markets (e.g., high frequency trading) and 
extremely large data systems (where the massive amount of data induces 
the large-scale property). Some of the problems that demand solutions 
within these domains are well known to this community, and include 
strategic planning, coordination, or task allocation. However, when 
small scale MAS solutions (and so called toy problems) are ported to 
large-scale MAS they typically suffer from scaling issues, unreliability 
or their solutions need assumptions whose results are inconclusive at 
best. Such issues arise due to the large number of participating agents 
(e.g., hundreds of thousands of homeowners in smart grids, or 
crowd-workers in crowdsourcing systems), or the large amount of data 
agents have to deal with (e.g., real-time trading data in high frequency 
trading, or large amount of entries in databases). Currently, such 
challenges are typically being addressed by the community either by 
limiting the agents' reasoning capabilities (assuming bounded 
rationality, suboptimal planning or limited information) or by limiting 
the agents' influence to local neighborhoods, significantly decreasing 
their action space. However, in this new ever-connected world, many 
applications, such as smart grids, or high frequency trading systems, 
require full agent models integrated at big scales that can exert their 
influence more widely. Thus, existing techniques, that rely on limiting 
either the agents' reasoning or action space, will fail in efficiently 
tackling the problems that occur in the abovementioned systems at scale.

To this end, this workshop aims to examine the major challenges induced 
by bringing together a nascent community of researchers from different 
research areas to genuinely discuss how to tackle these problems within 
large-scale autonomous agents and multi-agent systems. In particular, we 
aim to focus on addressing the challenges of adapting existing agent 
technologies to large-scale systems, and on novel approaches and 
solutions, that provide new, more efficient ways to describe the 
behaviour of agent-based systems at scale. These technical issues and 
the key discussion points to be addressed by researchers, include, but 
are not limited to, the following:

Novel models:
     Meta-agent based agent models: models with meta-agents that act on 
the behalf of groups of agents
     Community-based agent models: agent-type clustering and bootstrapping
     Dynamic, open system models: models with dynamically changing agent 
populations and behaviours
Computational aspects:
     Large-scale coordination & planning
     Large-scale single and multi-agent learning
     New performance metrics, analysis, and evaluation framework for 
massive MAS algorithms
Simulations:
     Large-scale simulation frameworks
     Real-time simulation at scale
     Benchmarking existing massive MAS techniques
Applications:
     Large-scale smart grids and energy systems
     Agent technologies for Big Data
     Large-scale crowdsourcing systems
     Large-scale voting systems
     Complex financial systems & high frequency trading
     Agent technologies for security
     Agent technologies for disaster management
     Agent technologies for social networks

*Important dates and deadlines

Deadline for the submission of full papers: February 11
Notification of acceptance/rejection: March 10
Deadline for the receipt of camera-ready papers: March 19
Workshop: May 4 or 5

*Submission instructions

Authors should submit full papers electronically in PS or PDF format on
EasyChair (https://easychair.org/conferences/?conf=massivemas2015).

Papers must be written in English, with a maximum length of 14 pages
(excluding references). Please format papers according to the Springer
LNCS Style (http://www.springer.com/computer/lncs?SGWID=0-164-6-793341-0).

Templates for Word and Latex are available. The receipt of 
submissionswill be acknowledged by email.
Submitted papers will be reviewed by the program committee.

Note: Workshop attendees need not register for the main AAMAS 
conference,but are encouraged to do so.

*Publication

TBA

*Program Committee

TBC

*Workshop Organizers

Enrique Munoz de Cote
INAOE, Mexico

Matthijs Spaan
Delft University of Technology, The Netherlands

Long Tran-Thanh
University of Southampton, UK

Matteo Venanzi
University of Southampton, UK

*Contact

Please contact Long Tran-Thanh (l.tran-thanh at soton dot ac dot uk) 
with any
inquiries.