We are already very much looking forward to the OR2017 International Conference on Operations Research which is taking place on September 06-08 in Berlin, Germany.
We will be giving several workshops and talks during the conference and will also have a booth in the exhibit area (Henry Ford Building, First Floor). If you are also coming to the conference make sure to come by and say Hi!
GAMS Pre-Conference Workshops
September 05, 13.00 / HFB|C
Part I: An Introduction (13.00 – 14.00)
We start with the basics of GAMS to develop algebraic models, solve them using state-of-the art algorithms, and introduce the key concepts of GAMS and the fundamentals of the language (e.g. sets, data, variables, equations). We’ll explore some case studies drawn directly from GAMS users in different fields.The largest part of the workshop will consist of an demonstration, where we are going to build a simple optimization based decision support application from scratch. We show how GAMS supports an easy growth path to larger and more sophisticated models, promotes speed and reliability during in the development phase of optimization models, and provides access to all of the most powerful large-scale solver packages. Along the way we will look at some of the data management tools included in the GAMS system and show how to analyze and debug large problems using the various tools available within GAMS.
This workshop requires no or limited knowledge about GAMS.
Break (14.00 – 14.30)
Part II: Advanced Hands-On Workshop (14.30 – 16.30)
This hands-on workshop, which is catered to the requirements of advanced GAMS users and application builders provides a great opportunity to learn about selected topics and best practices like:
- Use of the GAMS Object-Oriented API's to integrate GAMS models into different environments, like C#, Python, and Java.
- Code embedding in GAMS
- Stochastic programming in GAMS
These sessions are hands-on workshops, so please bring your laptop.
Talks during the conference
A distributed Optimization Bot/Agent Application Framework for GAMS Models
by Franz Nelissen
Thursday (Sep. 08) 14:45-16:15 [TD-02] / WGS|102
In this talk we will present a prototype of an optimization bot/agent application framework for GAMS models based on the GAMS .NET API. After the introduction we will show the basic architecture, features and the different elements of the system and illustrate them through an example.
Embedded Code in GAMS – Using Python as an Example
by Lutz Westermann
Wednesday (Sep. 07), 13:30-15:00 [WC-02] / WGS|102
This talk is about a recent extension of the General Algebraic Modeling System (GAMS): The “Embedded Code” facility. GAMS uses relational data tables as a basic data structure. With these, GAMS code for parallel assignment and equation definition is compact, elegant, and efficient. However, traditional data structure (arrays, lists, dictionaries, trees, graphs, …) are not natively or easily available in GAMS. Though it is possible to represent such data structures in GAMS, the GAMS code can easily become unwieldy, obfuscating, or inefficient. Also, in the past it was not easy to connect libraries for special algorithms (e.g. graph algorithms, matrix operations, …) to GAMS without some data API knowledge (GDX) and some disk I/O. To overcome these issues, the Embedded Code facility was introduced recently. It allows the use of external code (e.g. Python) during compile and execution time. GAMS symbols are shared with the external code, so no communication via disk is necessary. It provides a documented API and additional source code for common tasks so that the user can concentrate on the task at hand (e.g. computing shortest paths) and not the mechanics of moving data in and out of GAMS.
High Performance Computing with GAMS
by Fred Fiand and Michael Bussieck
Thursday (Sep. 08), 11:00-12:30 [TB-02] / WGS|102
BEAM-ME is a project funded by the German Federal Ministry for Economic Affairs and Energy and addresses the need for new and improved solution approaches for energy system models of vast size. The project unites various partners with complementary expertise from the fields of algorithms, computing and application development. The main focus is on large-scale linear programs (LPs) arising from energy system models. Such models have a block structure that is not well exploited by state-of-the-art LP solvers. Without considering this structure the models become quickly computationally intractable. Within the BEAM-ME project, new solution algorithms that exploit the block structure and utilize the power of High Performance Computers (HPC) are developed and will be made available to energy system modelers. Automatic detection of block structures in models has its limits and hence the user needs to provide block structure information via some model annotation. GAMS for some time has facilities to annotate a model. We discuss some extensions to the GAMS language which forms the interface between the energy system modeler and the newly developed algorithms. We provide an overview on the large variety of challenges we are facing within this project, present current solution approaches and provide first results.
Exam scheduling at United States Military Academy West Point
by Frederik Proske and Robin Schuchmann
Friday (Sep. 09), 9:00-10:30 [FA-02] / WGS|102
Each term the United States Military Academy (USMA) needs to schedule their exams. About 4000 cadets taking 5 - 8 exams each, need to be placed in 11 exam periods subject to several soft and hard constraints. Due to the short time frame in which the exams take place, a feasible solution in the sense that no cadet takes more than one exam per period cannot be obtained with a single exam version per course. So called makeups, alternative exams in another period, solve this problem. In order to reduce the extra work for instructors that must prepare those alternative exams, the number of makeups should be minimized. Makeups are also used to improve secondary objectives which occur at USMA West Point like the number of consecutive exams per cadet or to avoid that cadets take exams in certain periods (e.g. because of sport events they should attend). We consider two solution approaches: an integer programming approach using decomposition strategies and a nonlinear approach based on LocalSolver. We report numerical experiments for both methods based on real world data from the USMA West Point.