Summarizing the Metaheuristics International Conference 2013 in three words

This year’s Metaheuristics International Conference (MIC2013) has been held in Singapore. Around 70 presentations were made during the conference. In this post, I provide a short overview of what has been discussed through three keywords: hybridization, extensions and applications.

Hybridization

A lot of hybrid methods were proposed at this year’s MIC. A Hybrid is a method combining the features of two or more optimization algorithms. Researchers put a huge amount of effort on features that exploit specific structure of problems and on combinations of techniques that have proven to be successful. Judging by the reactions of presenters and attendees, efficiency (speed and solution quality) are the ultimate performance metrics for judging algorithms, rather than elegance or simplicity. Effective hybridization is one of the community’s areas of focus, judging from what has been submitted this year.

Extensions

Continuity plays a large role in what has been presented at MIC2013. Several authors proposed either faster algorithms for existing models or proposed extensions to existing models – along with a metaheuristic to solve it.  This is especially the case for the class of routing problems, for which many variants were presented.

It is rather natural for authors to propose extend their scientific contribution through either improved algorithms or additional constraints or variables to existing model. However, some of these talks unfortunately run low on context, assuming that you are familiar with the relevant literature and putting a focus on subtle nuances between the existing state-of-the-art and their method. As such, the value of attending these talks depends on whether you are familiar with the literature in the field and the authors’ previous version of the model/algorithm.

Applications

This term has been widely used – and sometimes abused – so as to cover almost any aspect. I prefer to limit the scope of what I call applications to cases when a metaheuristic has been used in the context of supporting decision-making with real users or to help scientists from other fields (such as biology or engineering) solve their domain-related problems.

Applications provide interesting stories to tell students about successes and hardships in applying OR in many situations. In these kind of talks, I prefer when the presenter discusses both what worked and what were the hardships encountered along the way. No applied OR projects go straight from A to B without a few unexpected twists, and I believe there is a lot to learn from one another in this regard. The best talk in this category was probably Mike Trick’s on Sports Scheduling, where he shared both sides of the story.

Overall, I enjoyed attending the MIC 2013 conference and I brought up quite a few interesting ideas to test at home. See you in Tunisia in 2015 (hopefully!)

Speak Your Mind

*