Venue: Industriens Hus, Copenhagen — May 8th, 2017
Registration: is now open
|09:00-10:00||Registration and refreshments|
|10:05-10:50||Michael Berliner Pedersen, Principal Analytics Consultant at SAS
Application of discrete event simulation to smart meter deployment
The Department of Energy and Climate Change in the UK has mandated the big six utilities in deploying smart meters to all their customers before 2020. SAS helped one on the big six in developing a smart meter deployment solution to manage the deployment planning. A component of this solution was a discrete event simulation (DES) model developed to investigate the operations of meter installations based on limited data availability before scaling up the deployment. The presentation will discuss the development of the DES model and its applications in the smart meter deployment process.
|10:55-11:40||Thomas Stützle, Senior Research Associate, Université Libre de Bruxelles
Automated Algorithm Design: Automatically generating high-performance algorithms
The design of algorithms for computationally hard problems is time-consuming and difficult for a number of reasons such as the complexity of such problems, the large number of degrees of freedom in algorithm design and the setting of numerical parameters, and the difficulties of algorithm analysis due to heuristic biases and stochasticity. In recent years, automatic algorithm configuration methods have been developed to effectively search large and diverse parameter spaces; these methods have been shown to be able to identify superior algorithm designs and to find performance improving parameter settings.In this talk, I will highlight the advantages of addressing algorithm design and configuration by algorithmic techniques; describe the main existing automatic algorithm configuration techniques; and discuss
various successful applications of automatic algorithm configuration. In particular, I will show how flexible algorithm frameworks can support the automatic design of high-performing hybrid stochastic local search algorithms. In fact, even for problems that have received very high attention in the literature new state-of-the-art algorithms can be obtained automatically, that is, without manual algorithm tuning. I will conclude arguing that automatic algorithm configuration has the potential to transform the way algorithms for difficult problems are designed and developed in the future.
|11:45-13:00||Lunch, provided by Industriens Hus|
|13:00-13:45||Anjana Rajan, Deployment Strategist at Palantir Technologies
Forward-Deployed Optimization: How Palantir uses data integration and human-computer interaction to solve complex problems in the field
Palantir is a private software company founded in 2004 in Palo Alto, California with the mission of helping the world’s most important institutions solve their most challenging and existential problems. Today, Palantir deploys software to help integrate and analyze every variety of data to many of the world’s largest institutions, including Fortune 100 companies, government organizations throughout the United States and Europe, and non-profits around the globe. The common thread across all of Palantir’s partnerships is organizations awash with data but lacking both the software to coherently integrate that data in one place and the software to help analysts answer the full range of questions they wish to ask of that data.
This presentation will discuss Palantir’s approach to data integration and human-computer interaction, and how that approach has informed work with customers for whom operations efficiency is a major, existential problem. In particular, we will discuss this through the lens of a product tailored for a few different types of operations analysts and engineers. Collectively, these users are responsible for optimizing the process of energy exploration based on deeply diverse and extremely high-scale data from a number of different sources.
|13:50-14:35||Niamh O’Connell, Data Scientist at Maersk Analytics
Optimisation of Fuel Consumption in Liner Shipping Schedule Design
Maersk Transport and Logistics, in collaboration with DTU Management Engineering, are addressing the challenge of minimising fuel costs on our liner shipping network. Fuel costs represent a significant portion of the complete costs portfolio at Maersk, and we are pursuing the development of this optimisation as part of an overall drive towards the integration of data-driven analytics into our traditional business model. In this talk, I will describe the mathematical constructs necessary to represent the transportation of shipping containers through a liner shipping network, and the constraints employed to represent the operational and commercial considerations when designing a network schedule.
|14:40-15:10||Coffee & networking|
|15:10-15:55||Rune Møller Jensen, CEO at Optivation
Stowage Planning Optimization: The Last Mile
Mckinsey has recently estimated a yearly global loss of $2B due to poor stowage planning of container vessels. Optivation develops stowage planning optimization algorithms for the StowMan and Xvela software products of NAVIS. Over the last four years, we have experienced the true length of the “last mile”. We have been challenged by the problem’s high dimensionality, complex business requirements, and the fact that it is an art rather than a science. In my talk, I would like to describe the issues, we have been facing, and our approach to solving them. My hope is to highlight general challenges of applying operations research.