Call for papers: WCCI 2014 Special Session "Learning and optimization in multi-criteria dynamic and uncertain environments"

Motivation

Many real-world environments have inherently multiple criteria (or objectives) that can be aligned as well as conflicting, resulting in complex Pareto. To efficiently explore these complex fronts, new exploration/exploitation techniques are needed. Inspiration for such techniques can be found in multi-objective optimization.

Usually, the state of the system is changing in time, making the environment dynamic. There are many interesting applications in the field of engineering, i.e. automatic control, robotics, where one wants to simultaneously fulfill different criteria using a number of constraints or preferences.

The task of an optimization algorithm in multi-criteria environments is to learn a strategy that optimizes all criteria at the same time or to find a good compromise solution. Thus, learning in the multi-criteria framework can be considerable harder than in the standard single objective framework. Currently, there are two major, conceptually different, approaches dealing with dynamic environments: i) Reinforcement Learning and ii) Evolutionary Algorithms.

Reinforcement learning is traditionally formalized within the Markov Decision Process (MDP) framework. An agent takes actions in a stochastic and possibly unknown environment, and moves between states in this environment. After each action, the agent receives a reward signal in order to develop a strategy that maximizes its long-term (cumulative) reward.

The approach of an Evolutionary Algorithms is to continuously track the optimum in dynamic environments, or to find a robust solution that operates optimally in the presence of uncertainties. This poses serious challenges to conventional EAs that are not conceptually designed to handle environmental changes.

Goal

The main goal of this special session is to start the process of unifying and streamlining research on learning in dynamic and uncertain multi-criteria environments which for time being seems to evolve independently and disconnected in Reinforcement learning and Evolutionary Algorithms. We want to bring together researchers from machine learning, optimization and artificial intelligence, interested multi-criteria decision making and/or multi-objective optimization in dynamic and uncertain environments. We also encourage submissions related to multi-criteria decision-making and/or multi-objective optimization in other areas such as operation research, games and real-world applications.

Ideally, the special session will bring together researches with different background in Machine Learning and Optimization in order to help them identifying common ground for their work.

Topics of interest

Topics of interests include but are not limited to:

• Multi-objective reinforcement learning
• Multi-objective optimization algorithms such as meta-heuristics, evolutionary algorithms, etc. for dynamic and uncertain environments
• Theoretical results on the learnability in multi-objective dynamic and uncertain environments
• Novel algorithmic frameworks for multi-objective environments 
• Multi-criteria aspects of robotics
• Multi-objective self-adapting systems
• Multi-objective automatic configuration systems
• Multi-objective games
• Multi-criteria decision making in dynamic and uncertain environments
• Real-world applications in engineering, business, computer science, biological sciences, scientific computation, etc. in Dynamic and Uncertain Environments
• Multi-criteria dynamic/reactive scheduling and planning

Information for Authors

This section is part of IEEE International Joint Conference on Neural Network 2014 (IEEE IJCNN 2014)

1)    Information on the format and templates for papers can be found here:
       http://www.ieee-wcci2014.org/Paper Submission.htm
2)    Papers should be submitted via the IJCNN 2014 paper submission site:
       http://ieee-cis.org/conferences/ijcnn2014/upload.php3
3)    Select the Special Session name in the Main Research topic dropdown list
4)    Fill out the input fields, upload the PDF file of your paper and finalize your submission by the deadline of December 20, 2013

Organizers

Dr. eng. Madalina M. Drugan,
Computational Modeling group, Artificial Intelligence Lab of Computer Science Department, Vrije Universiteit Brussels, Belgium
e-mail: Madalina.Drugan@vub.ac.be

Prof. dr. Ann Nowe,
Computational Modeling group, Artificial Intelligence Lab of Computer Science Department, Vrije Universiteit Brussels, Belgium
e-mail: Ann.Nowe@vub.ac.be

Important dates

Paper submission: 20 December, 2013
Decision: 15 March, 2014
Final paper submission: 15 April, 2014
Conference dates: 6-11 July, 2014