Building a platform for modeling and solving Optimization problems (Master seminar)

Having such an extensive use, there are already a plethora of tools and software suites out there, open and closed source, free and paid, that enable scientists, mathematicians and business analysts to solve complex problems in their everyday work (Guobi, IBM Cplex, FICO). These tools have developed for many years now and yet some of them fail to provide modularity, good usability and extensibility of their software, while also having unaffordable prices for many.

We plan to build a multi-user platform, available as a web app online, for modelling and solving linear & mixed-integer linear optimization problems. This software suite will use state-of-the-art libraries and technologies in order to achieve performance, accuracy and usability. Moreover, this platform will enable users to interact with each other by sharing their work in order to reuse models, avoid mistakes or bad formulations and get feedback for their work.

With these in mind, the motivation of this thesis is to understand problems with existing solvers, research the latest technologies for the software set we are planning to build and lastly, to go hands-on with implementing these technologies. During this presentation I will explain what the problem is, what are the existing software solutions out there (and their drawbacks), what will the architecture of the new platform be, my motivations behind certain technologies I will use during implementation, how do I plan to carry out the implementation step by step, and also some first steps of work done during master seminar time.

-----------------
Join Zoom Meeting
Meeting ID: 527 278 8807

Note: for people outside D1 interested in listening to this talk, please contact Sándor Kisfaludi-Bak at skisfalu@mpi-inf.mpg.de for the password.