ECiP @GECCOConf 2024

July 16, 2024 (hybrid)

Evolutionary Computation in Practice

In the Evolutionary Computation in Practice (ECiP) track, well-known speakers with outstanding reputation in academia and industry present background and insider information on how to establish reliable cooperation with industrial partners. They actually run companies or are involved in cooperation between academia and industry. 
If you attend, you will learn multiple ways to extend EC practice beyond the approaches found in textbooks. Experts in real-world optimization with decades of experience share their approaches to creating successful projects for real-world clients. Some of what they do is based on sound project management principles, and some is specific to our type of optimization projects. If you are working in academia and are interested in managing industrial projects, you will receive valuable hints for your own research projects.
In 2024, ECiP will be a hybrid event. We will do our best to enable opportunities for establishing contacts among participants.

Program (Overview)

Day: July 16th, 2024

Time: 11:00 – 13:00 (UTC+10)

Time	Speaker	Title	Affiliation
11:00 – 11:25	Carlos M. Fonseca & Bogdan Filipič	The Randomised Optimisation Algorithms Research Network: Objectives and Opportunities	University of Coimbra, Portugal & Jožef Stefan Institute, Slovenia
11:25 – 11:50	Simon Ratcliffe	Six million compute minutes solving billion dollar problems – a successful GA deployed industrially.	University of Adelaide, Australia
11:50 – 12:15	Steffen Limmer	Automated Machine Learning	Honda Research Institute EU
12:15 – 12:40	Thomas Bartz-Beielstein	Simplifying Hyperparameter Tuning for Industrial Applications with spotPython: Examples from PyTorch, Scikit-Learn, and River	University of Applied Sciences – TH Köln, Germany

Carlos M. Fonseca & Bogdan Filipič:
The Randomised Optimisation Algorithms Research Network: Objectives and Opportunities

Abstract

The Randomised Optimisation Algorithms Research Network (ROAR-NET) is a research networking Action funded by COST (European Cooperation in Science and Technology) that currently involves over 300 members from 35 COST countries and 7 other countries including Australia, Mexico, South Africa and the US. Running from October 2023 to September 2027, the network “aims at making randomised optimisation algorithms widely competitive in practice by identifying and reducing obstacles to their adoption at the scientific, technical, economic, and human levels.” To this end, it is structured as six Working Groups focusing on problem modelling and user experience, mixed continuous and discrete optimisation, single- and multiobjective optimisation, optimisation under uncertainty, algorithm selection and configuration, and benchmarking. ROAR-NET organises on-site meetings, Workshops, and Training Schools, as well as online activities. It also supports Short-Term Scientific Missions (STSMs), which are visits by researchers and innovators to host organisations located in a different country than their country of affiliation, including non-COST countries, for specific work to be carried out. The presentation summarises the objectives and expected outcomes of ROAR-NET. It also presents the opportunities afforded by the network and shows how both academia and industry can get involved.

This presentation is based upon work from COST Action Randomised Optimisation Algorithms Research Network (ROAR-NET), CA22137, supported by COST (European Cooperation in Science and Technology).

Short CV: Carlos M. Fonseca

Carlos M. Fonseca is an Associate Professor at the Department of Informatics Engineering of the University of Coimbra, Portugal, and a member of the Adaptive Computation group of CISUC, the Centre for Informatics and Systems of the University of Coimbra. He obtained his doctoral degree from the University of Sheffield, U.K., where he also conducted post-doctoral research. His main research interests are in multiobjective optimisation, evolutionary computation, experimental evaluation of optimisation algorithms, and practical applications of optimisation. He currently focuses mainly on preference articulation in multiobjective optimisation and on the computational modelling of optimisation problems. He was General or Technical co-Chair of major international conferences, a mentor at several Summer Schools, and Guest co-Editor of three journal Special Issues (one to appear). He was Working Group Leader in COST Action ImAppNIO (CA15140) and now serves as Action Chair and Grant Holder Scientific Representative in COST Action ROAR-NET.

Short CV: Bogdan Filipič

Bogdan Filipič is a senior researcher and head of Computational Intelligence Group at the Department of Intelligent Systems of the Jožef Stefan Institute, Ljubljana, Slovenia, and professor of Computer Science at the Jožef Stefan International Postgraduate School. He received his PhD degree in Computer Science from the University of Ljubljana. His research interests are in artificial intelligence, evolutionary computation and randomised optimisation. He currently focuses on evolutionary multiobjective optimisation where he studies solution visualisation, constraint handling, and problem characterisation. He was a guest lecturer at the University of Oulu, Finland, and the VU University Amsterdam, The Netherlands, gave tutorials at recent CEC and GECCO conferences, and was General Chair of BIOMA and PPSN conferences. He is also active in applying evolutionary computation in practice and has led optimisation projects in engineering design, manufacturing, and energy management. He was Working Group leader in COST Action 526 APOMAT and now serves as Grant Awarding Coordinator in COST Action ROAR-NET.

Simon Timothy Ratcliffe:
Six million compute minutes solving billion dollar problems – a successful GA deployed industrially

Abstract

The location, amount and order in which ore is mined from the earth is among the most important decisions a mining enterprise must make to determine profitability. Many linear models for this problem have been posed and the optimisation of them in open cut mining is a well studied problem. Non-linear formulations make for better models in many situations and the optimisation of these are far less studied. This talk presents such a formulation and how a simple GA has been engineered and deployed in commercial software to enable miners to make better decisions. The problem formulation will be presented and the solution approach will be discussed along with case studies.

Short CV

Simon Ratcliffe is an experienced company executive, team builder, software engineer and product development entrepreneur. During a long career in the private sector, he has been instrumental in the design, development and commercialisation of several leading software and hardware products that continue to enjoy success in the highly competitive global mining technology industry. He continues in industry practice today as Head of Experimentation and Technology at Maptek as well as holding a Professorship at the University of Adelaide in Software Engineering. Simon’s research interests are in computational geometry, high performance numerical computing, evolutionary computation and machine learning – particularly where and how they intersect with industry challenges and commercial opportunities.

Steffen Limmer:
Automated Machine Learning

Abstract

The manual development of well performing machine learning pipelines is a time-consuming task, which requires a substantial amount of human expertise. The automation of this task, also known as  automated machine learning (AutoML), is gaining increasing attention in academia as well as industry. This talk gives a practical introduction to AutoML and to TPOT, one of the most popular AutoML tools. Furthermore, an overview to AutoML related activities of the Honda Research Institute Europe is provided.

Short CV

Steffen Limmer received the M.Sc. degree (Diploma) in computer science from the University of Jena, Germany, in 2009, and the Ph.D. degree in engineering from the University of Erlangen-Nürnberg, Germany, in 2016. He worked as a Scientific Assistant at the Chair of Computer Architecture, University of Erlangen-Nürnberg, from 2009 to 2016.  Since December 2016, he has been a Senior Scientist at the Honda Research Institute Europe. His current research interest includes optimization and data-driven modeling in the context of energy management systems.

Details

Organizers

• Thomas Bartz-Beielstein, IDE+A, TH Köln, Germany
• Richard Schulz, IDE+A, TH Köln, Germany
• Danial Yazdani, University of Technology Sydney, Australia

About GECCO

GECCO Schedule at a Glance (https://gecco-2024.sigevo.org/Program)