Write up about the Conference
The 1st International Conference on Engineering and Applied Sciences Optimization (OPTI-i) was held at the Kos Island Convention Center, Kos Island Greece, last June 4-6, 2014. It was organized by the Department of Transportation Planning and Engineering and the Department of Structural Engineering of the School of Civil Engineering of the National Technical University of Athens. This conference was endorsed by the International Society for Structural and Multidiciplinary Optimization (ISSMO), Transportation Research Board (TRB), European Community on Computational Methods in Applied Sciences (ECCOMAS), Hellenic Institute of Transportation Engineers (HITE), Greek Association for Computational Mechanic (GRACM) and John Argyris Foundation (JAF). The OPTI-i conference facilitates the exchange of ideas in topics of mutual interests and is serving as a platform for establishing links between research groups with complementary activities. Thus, the communities of Mathematical Programming and Nature-Inspired Search Algorithms become more familiar with important application areas with great social impact, arising from real-world problems. The Technical Programme consists of six (6) plenary lectures, twelve (12) semi-plenary lectures, nine (9) keynote lectures, and 20 mini-symposia with more than 400 paper presentations in total. Lectures included topics in Topology Optimization and Additive Manufacturing; Liner Containership Routing and Scheduling; and Simulation and Optimization in Transportation; Computational Design of Materials and Multi-scale Structure; Adjoint Methods for Steady & Unsteady Optimization in Fluids; Multiple Criteria Decision Making and Optimization; Location, Logistics, Transportation, Traffic and Telecommunication-Networks, Optimization Software; Hybrid Optimization Algorithms; Non-smooth Numerics via Piecewise Linearization; Metaheuristics Algorithms for Combinatorial problems and Engineering Applications; Shape and Topology Optimization; Artificial Intelligence and Soft computing Methods for Transportation & Infrastructure Engineering Problems; Structural Optimization in Composite Materials, Transient Load and Crash, Fatigue and Fracture; and OR in Health, Life Sciences, Industry and Natural Resources.
Feedback on paper presented
There were three questions and one suggestion raised:
1) How did you measure good fit of the models obtained by the proposed hybrid algorithms and why is DESA the best algorithm? Answer: Mean of Square Error (MSE) was used to evaluate and compare Differential Evolution alone, Simulated Annealing alone, DESA-the hybrid algorithm. In our experiment DESA was found to have the smallest MSE compared to those of DE, SA and even to MLE and OLS.
2) Why did you use MSE while the results were not really normally distributed across all values of parameters? Answer: I used MSE to evaluate and compare the performance of the five algorithms, since I wanted to show that DESA is generally efficient in fitting AR(1) model with parameter values from 0.1 to 0.9, and to show that DESA also performs very well even on nearly non-stationary AR1 processes, unlike the other optimization algorithms.
3) Did you consider and compare the computer times? Yes. In our preliminary experiments, we have observed the SA is very slow compared to DE which is very fast. So we came to the hypothesis that combining DE with SA will provide better solutions and faster convergence. And indeed we obtained smaller MSEs and DE improved computer time of SA.
Suggestion: To extend the DESA algorithms to other time series fitting problems and to practical data applications.
Future directions of research presented
The paper presented can be extended to a problem of Fitting ARIMA/ VARIMA models (univariate and multivariate) to time series data and its application to a longer time series. Moreover, the paper can be extended to the Generalized Models for Spectral Matrices and its application to EEG data.
Potential foreign collaborators
Michael Kokkolaras, McGill University, Canada
“Direct Search and Statistical Surrogates for Simulation-based Design Optimization”
Jacques Francis Periaux, CIMNE Barcelona, Spain Univ. Jyvaskyla, Finland
“Multi Hybridization Techniques for Advance Parallel Evoutionary Design in Aerospace and Structure Engineering”.
Kyriakos C. Giannakoglou,
Professor NTUA, Lab. of Thermal Turbomachines,
School of Mechanical Engineering, National Technical University of Athens,
e-mail: [email protected]
Other important contacts and insights
- Manolis Papadrakakis, National Technical University of Athens, Greece
- Matthew G. Karlaftis, National Technical University of Athens, Greece
- Nikos D. Lagaros, National Technical University of Athens, Greece
Institute of Structural Analysis and Antiseismic Research
National Technical University of Athens
Zografou Campus, Athens 15780, Greece
INSIGHTS: Time was so short but I had a great time. I have shared and learned a lot from attending the conference, and the same time I have treated myself a break from my never ending review of statistics and probability concepts. At this point, I would like to thank the University of the Philippines for providing financial support through the Office of the Vice President of Academic Affairs- Research Dissemination Grant (OVPAA-RDG). The on-time approval of my application to the travel grant gave me the courage to register to the conference. I also took the risk of buying roundtrip air-tickets ahead of time; since, I was so confident that my application for additional travel grant will be approved. Taking all the risks, I have forced myself to write and finish the full paper which is a requirement of the application to the Commission on Higher Education (CHED) Support for Paper Presentation in International Conferences. My struggles in availing financial supports, registering to the conference, obtaining Schengen Visa, spending own money for buying tickets, and the thought of travelling alone gave a lot of stress. Nonetheless, all of these made me even stronger to face the life of an academician, here at our prestigious institution, The University of the Philippines.
Short write-up of one’s participation (to be used to feature/publicize the grantee’s participation in the conference)
The conference days which usually started 8:30am and ended 6:30pm were very long and tiring. However, the topics were very useful in my scientific endeavor of additional knowledge on how to do research in collaboration with other disciplines. With all the papers presented, I became more familiar with important applications in structural, aero-space, mechanical, civil, chemical, naval and bio-engineering to applied problems in economics, physics and biology. In addition, I was exposed to advance computational methods and software tools which can highly assist in handling combinatorial optimization to continue facilitate the exchange of ideas in topics of mutual interest and serve as a platform for establishing links between research groups with complementary activities.