JOINT Workshop on

UHEUR – Understanding Heuristics: How do we get the best of both theory and empirical methods?

AND

PARCO – Parallel Cooperative Search Methods

Held in conjunction with the

11th International Conference on Parallel Problem Solving From Nature (PPSN 2010)

UHEUR Organisers Ender Özcan^1, Andrew J. Parkes^1 and Jon Rowe^2 1. [>>] Automated Scheduling, Optimisation and Planning (ASAP) Research Group [>>] School of Computer Science [>>] University of Nottingham, UK {exo, ajp} AT cs.nott.ac.uk 2.  [>>] School of Computer Science      [>>] University of Birmingham, UK J.E.Rowe AT cs.bham.ac.uk

PARCO Organisers *Djamila Ouelhadj, **Ender Özcan, and ***Michel Toulouse *Logistics and Mathematics> Management Group (LMMG), Department of Mathematics, University of Portsmouth, UK. Email: djamila.ouelhadj AT port.ac.uk **Automated Scheduling, Optimisation and Planning (ASAP) Research Group, Computer Science,University of Nottingham, UK. Email: exo AT cs.nott.ac.uk ***Department of Computer Science, Oklahoma State University, USA. Email: michel.toulouse AT cs.okstate.edu

---

Workshop Program

Please note that these workshops have been combined and rescheduled.

Time	Speaker/Authors	Title	Room
14:00 - 14:10	Andrew Parkes	Introduction to the workshop and brief description of the LANCS initiative (which gave some financial support), and in particular the Heuristic Understanding research cluster	Aula U-2
14:10 - 15:00	El-Ghazali Talbi	A unified view of parallel metaheuristics	Aula U-2
15:00 - 15:30	J.L.J. Laredo, J.J. Merelo, C. Fernandes, A. Mora, M.G. Arenas, P.A. Castillo and P. Garcia-Sanchez	Influence of the Population Structure on the Performance of an Agent-based Evolutionary Algorithm	Aula U-2
15:30 – 16:00	Coffee Break
16:00 – 16:45	Andrew Parkes	Phase Transitions	Aula U-2
16:45 – 17:15	B.Bilgin, P. De Causmaecker, S. Haspeslagh, T. Messelis, G. V. Berghe	Hardness studies for nurse rostering problems	Aula U-2

---

Overall Description

In this combined workshop we focus on challenges to heuristics arising from parallel implementations, and also the problems of creating and tuning heuristics on real problems.

---

UHEUR Description Whether (meta)-heuristics are implemented sequentially or parallel they often require significant tuning to work effectively. Often this tuning is without any a priori assumptions of how control parameters might depend on features of the problem. However, this can require a lot of computational effort and also has the danger of being limited to only problems that it has been seen before. In particular, we often would like to be able to tune of small instances (for speed) but then adjust so as to be able to run on large instances. This will require some theory of how such extrapolation in size, or outside of the features of the training set, out to be performed. An anology is the difference between non-parametric and parametric statistics; the former is does not assume any underlying probability distribution and the latter can (for example) assume a Gaussian. Naturally, the latter can give stronger results and with smaller sized samples. Hence, we might call this "Parametric Parameter Tuning". Of course, this is a challenging problem; but we hope to be able to discuss any existing work and how the community might meet the challenge.

PARCO Description Interest in cooperative search has risen considerably due to its success to provide novel ways to combine different search algorithms. Current research has shown that cooperative search which involves the parallel execution and cooperation of several (meta/hyper-) heuristics and/or exact methods can improve the performance of the algorithms, especially when dealing with large scale combinatorial optimisation problems. Moreover, parallel cooperative search can be used to provide more powerful problem solving environments in a variety of fields, ranging, for example, from finance to bio- and health-informatics. The goal of the workshop is to bring international researchers and practitioners together to cover the most recent trends in this emerging area, complementing it with practical implementations and demonstrations in this field.

See [>>] for the original description of the workshop

See [>>] for detailed description

[Top]