Sign Out
Logged In:
 
 
 
 
 
Tab Image

OR56 - Making an Impact Workshops

Making an Impact: In Practice

OR56, Royal Holloway University, 10 September 2014

Workshops, tutorials and technique tasters

In the afternoon, a series of sessions will be run by OR members who are experts in their fields. These sessions will give participants the opportunity to learn, collaborate and debate on a wide range of topics to count towards their CPD. The workshops will be arranged into two back-to-back sessions and participants will be able to sign up to the sessions on arrival Wednesday morning 10 September. The sessions are first come first served and under-subscribed sessions may be cancelled.

The current list of confirmed workshop titles and abstracts are below, more are to come:

Get me out of this grid! agent-based modelling in geospatial environments
Benjamin Schumann
(DecisionLab)
Most people like beautiful maps. They provide a huge amount of information visually without overloading our limited information processing capabilities. Moreover, many OR practitioners also like agent-based modelling: it is a useful alternative for solving many OR problems by defining individual behaviour. The problem is that these two worlds rarely meet. They should! Often, OR problems require agents to act in a geospatial environment: Where do you place water reservoirs? What airport destinations are best suited for an airline? How do people use road networks? Geospatial analysis can provide new insights into problems as maps contain a lot of information without overwhelming users.  To date, bringing together Agents and geospatial analysis was mostly handled in two ways:
• Neglect the geospatial bit altogether, or
• Draw geospatial “map” manually.
Why is that? Quite simply, both agent-based modelling and geospatial analysis tools are rather sophisticated. Mastering both is a challenge. Linking both worlds is even harder. We at decisionLab do our best using the AnyLogic software. So far, AnyLogic has been a great tool for agents but was weak with maps. This is about to change...
In this workshop, you will get to see the future of agents and geospatial modelling: you will send taxi drivers through the maze of central London. You will create a network of pharmacies in one click and get delivery trucks to serve them. You will see agents actually following Google Maps routes (or OpenStreetMap, if you prefer)! And you will learn about other cool capabilities of agents in spatial environments that were fiendishly difficult to do until ... now!

Optimising the real world, robustly
Andy Harrison (FICO)
We all know that the world is messy, complicated, complex, uncertain and non-linear.  But many areas of Operational Research have been very successful by assuming that the world is not messy but simple, certain and linear.  Linear programming, in particular, has delivered significant, tangible benefits to industrial problems by assuming the world is certain and linear.
Bigger and every more powerful computers, grids and clouds together with breakthroughs in solver technology have meant we can solve bigger and bigger linear programmes to allow us to tackle industrial problems that just a decade ago were impractical to even consider.  This will become even more so, as Big Data delivers tera-, peta- and perhaps even zeta-bytes of data to formulate ever larger linear programmes.
However, with ever more data comes more errors.  In addition predictions made from this data are often uncertain.  With this in mind, earlier this year, the FICO™ Xpress Optimization Suite introduced support for robust constraints.  This means it is now possible to model uncertainty and obtain robust optimal solutions efficiently.
In this workshop we will use the FICO™ Xpress Optimization Suite to explore the implementation of robust constraints and their impact on the solution.  This will include the implementation of a robust optimisation model for production planning under uncertainty.
So come to the workshop ready to do some hands on modelling.  If you want to bring your laptops, I can supply a copy of Xpress to use on a USB key (or if you’re keen you can get your own copy from http://www.fico.com/xpress).  But whether you have a laptop available or not you will have an opportunity to can roll up your sleeves and build some robust optimisation models.
If you enjoy the workshop and you want to explore optimisation and using the FICO™ Xpress Optimization Suite more, there is a three day training course organised through the OR Society, which will run 18th – 20th November in Birmingham.  For more details see http://www.theorsociety.com/Pages/Training/Courses/2014/324.aspx.

Introduction to Data Science
Sayara Beg
(Datanut)
The Data Science workshop will give attendees a taste of what a data scientist would need to think about when approaching a challenge. In the workshop we take a small problem and break it down step by step, in a fun way, following data science governance and standards. You will leave the workshop with an understanding of data science concepts.

The clash of theory and reality – when project management becomes people management
Sophie Carr (Bays Consulting), Ian Seath (Improvement Skills Consulting Ltd) and David Wrigley (ORvis Consulting Ltd)
In this interactive workshop we’ll examine three typical project scenarios where people, not data or methodologies, are at the heart of the problem, using real world case studies to identify how problems can be identified early and be avoided or addressed if they do occur.

Field notes from consultancy using SoftOR and problem structuring methods
Giles Hindle
(Hull University Business School)
SoftOR and Problem Structuring Methods constitute a set of approaches to tackling complex problems and strategic thinking. They have been successfully used in a wide range of contexts and usefully expand the area of application for analytical consultants and the OR discipline. Drawing on the author’s own Action Research programme spanning 20 years, a set of field notes for successful practice is presented. The author will explore typical problems, intervention modes and project designs and give practical advice on the application of techniques such as situation mapping, picturing, conceptual modelling and systems thinking. Reference is also made to a spectrum of consulting approaches from facilitative/ participative to analytical/ expert.

Creating value from Big Data and business analytics: the organisational challenge
Richard Vidgen and John Morton
(Hull University Business School and CPM Consulting)
Advances in ICT have led to an exponential growth in the volume, variety, granularity and rate of generation of the data (velocity) that organizations can access. Big Data and business analytics are becoming an increasing worry for the company board and senior executives. Have they “missed the boat”, is there any value in the data anyway, and if there is, is it relevant to their organization? A summary of the results of recent research into how organisations create value from [big] data will be presented. Attendees of the workshop will then be asked to engage in structured break-out groups to identify the challenges that organisations face in creating value from their data. The findings of each group will be summarized in plenary. If appropriate, the workshop will be followed up with a Delphi study to consolidate the issues and to identify priorities for action.

Model students: an operational researcher’s guide to upskilling for Analytics
Michael Mortenson
(Loughborough University)
With numerous reports forecasting significant skills shortages, and Harvard Business Review declaring “data scientist” as the sexiest job of the 21st Century, does big data mean a big opportunity or a big threat for operational researchers? As analytics grows in reach and prominence, how can OR maximise this opportunity and establish a position at the heart of analytics?

Incorporating the results of two years of research into the area, this talk will discuss both the “hard” and “soft” skills most in demand with employers, from programming languages to predictive modelling, and some of the low- and no-cost tools that can be used to “upskill”. The workshop will also provide the opportunity to share your experiences of advanced analytics in practice, and debate the future of OR in the era of big data.

Machine learning with R: an introduction
Richard Vidgen (Hull University Business School (HUBS))
“R is the world’s most powerful programming language for statistical computing, machine learning and graphics as well as a thriving global community of users, developers and contributors. R is free, open-source software distributed and maintained by the R-project.” (Revolution Analytics). R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering) and graphical techniques, and is highly extensible. In proprietary packages, such as SPSS, the user has limited ability to change the environment, instead relying on the algorithms developed for them by SPSS developers (and must also pay for this constrained problem-solving). As a long-time user of SPSS and structural equation modeling software (AMOS, Mplus, SmartPLS) I’ve recently begun to make greater use of R for predictive analytics. R has more than 5000 thousand packages available, free of charge, that allow for many types of data science techniques, such as Bayesian analysis, regression trees, cluster analysis, social network analysis, and text mining. There is quite a steep learning curve to working with R. In this workshop I will seek to signpost an accessible route to becoming an R user and illustrate this with a comparison of multiple regression in SPSS and in R.

To demonstrate further the flexibility and power of R we will access Facebook data. Attendees are encouraged to download and install R (http://www.r-project.org/index.html) and the development environment R Studio (http://www.rstudio.com/products/RStudio/) in advance of the workshop.