• GECCO 2015 workshop on automatically configurable algorithmic frameworks

Automatically Configurable Algorithmic Frameworks

This workshop aims at bringing together researchers, developers and end users of algorithmic frameworks for solving optimization problems. The focus of the workshop will be on the design, analysis and automatic configuration of such frameworks.

The proliferation of metaheuristics, including many types of local searches, evolutionary algorithms, and swarm intelligence approaches, has given rise to software frameworks that aim at generalizing and hybridizing diverse techniques by combining components from different optimization algorithms. Examples of such frameworks are PaGMO [1], ParadisEO [2], Shark [3], SATenstein [4], jMetal [5] and the MOACO framework [6]. These frameworks are usually characterized by a large number of parameters that allow selecting and configuring the behavior of specific algorithmic components. A parameter configuration is, hence, an instantiation of a particular optimization algorithm fine-tuned to a specific application scenario. Inevitably, the performance of the frameworks strongly depend on their parameter settings and manually tuning such parameters is a hard and tedious task for the algorithm designer. Moreover, the ability of the framework to generate novel algorithm designs will depend on its flexibility when combining algorithmic components. Automatic configuration techniques (parameter tuning) allow the algorithm designers to focus on the design of new algorithmic components, and on how to combine them into a flexible framework. The final algorithm design decisions will be taken by the automatic configuration technique that will search in the algorithmic design space defined by the framework for the best algorithm for the problem at hand [7,8,9].

The automatic configuration of algorithmic frameworks poses challenges in terms of generalization of algorithmic components, identification of key components to be tuned, algorithmic configuration methods, different types of abstraction (top-down versus bottom-up), and automatic design approaches.

We invite researchers to submit their original and unpublished works on the aforementioned challenges. In particular, we will welcome:

• Novel ideas on how to design, extend and analyze metaheuristics frameworks in order to enable automatic configuration and design.
• Proposals for new algorithmic frameworks designed to be automatically configured.
• Theoretical studies that propose generalized models of metaheuristics to be implemented in (new) flexible frameworks.
• Studies that analyze and evaluate the conceptual models underpinning existing frameworks
• Experimental studies that compare the capabilities of existing frameworks to produce state-of-the-art algorithms for challenging optimization problems in an automatic fashion.

References

1. F. Biscani, D. Izzo, and C. H. Yam, "A Global Optimisation Toolbox for Massively Parallel Engineering Optimisation," CoRR, vol. abs/1004.3824, 2010.
2. S. Cahon, N. Melab, and E. G. Talbi, "ParadisEO: A framework for the reusable design of parallel and distributed metaheuristics," Journal of Heuristics, vol. 10, no. 3, pp. 357–380, 2004.
3. C. Igel, V. Heidrich-Meisner, and T. Glasmachers, "Shark," The Journal of Machine Learning Research, vol. 9, pp. 993-996, 2008.
4. A. R. KhudaBukhsh, L. Xu, H. H. Hoos, and K. Leyton-Brown, "SATenstein: Automatically building local search SAT solvers from components," in Proceedings of IJCAI- 09, C. Boutilier, Ed. AAAI Press, Menlo Park, CA, 2009, pp. 517–524.
5. J.J. Durillo, A.J. Nebro, "jMetal: a Java Framework for Multi-Objective Optimization," Advances in Engineering Software 42 (2011) 760-771.
6. M. López-Ibáñez and T. Stützle, "The automatic design of multi-objective ant colony optimization algorithms," IEEE Transactions on Evolutionary Computation, vol. 16, no. 6, pp. 861–875, 2012.
7. H. H. Hoos, "Programming by Optimization," Communications of the ACM, vol. 55, no. 2, pp. 70-80, February 2012.
8. F. Mascia, M. López-Ibáñez, J. Dubois-Lacoste, and T. Stützle, "Grammar-based Generation of Stochastic Local Search Heuristics Through Automatic Algorithm Configuration Tools," Computers & Operations Research, vol. 51 no. 0, pp. 190-199, 2014.
9. M. E. Marmion, F. Mascia, M. López-Ibáñez, and T. Stützle, "Towards the Automatic Design of Metaheuristics," In H. C. Lau, et al., editors, Proceedings of 10th Metaheuristic International Conference (MIC 2013), Singapore, August 5-8, 2013, pp. 215-217, 2013.

Submission Details

We invite submissions of papers in PDF format, either short-papers (4 pages) or full papers (between 6 and 8 pages), and formatted following the preparation instructions for GECCO 2015. The review process of the workshop is not double-blind, thus papers should not be anonymous but contain full authorship details.

Papers should be submitted directly to fmascia :at: ulb.ac.be and manuel.lopez-ibanez :at: ulb.ac.be. All accepted papers will be presented at the workshop and appear in the GECCO Conference Companion Proceedings published by ACM.

Important Dates

Organisers

Manuel López-Ibáñez Dr. Manuel López-Ibáñez is a postdoctoral researcher (Chargé de recherche) of the Belgian Fund for Scientific Research (F.R.S.-FNRS) working at the IRIDIA laboratory of Université libre de Bruxelles, Brussels, Belgium. He received the M.S. degree in computer science from the University of Granada, Granada, Spain, in 2004, and the Ph.D. degree from Edinburgh Napier University, U.K., in 2009. He has published 14 journal papers, 6 book chapters and 33 papers in peer-reviewed proceedings of international conferences on diverse topics such as evolutionary algorithms, ant colony optimization, multi-objective optimization, and various combinatorial and real-world optimization problems. His current research interests are experimental analysis, automatic configuration and automatic design of optimization algorithms, for single and multi-objective optimization. He is the lead developer and current maintainer of the irace automatic configuration method.