PuzzlOr : Electrifying
PuzzlOr problems are nice because they are simplified versions of real world problems of interest. Last December problem is a simplified version of an interesting logistics problem. A variety of method can be used to attack them, see for instance this interesting post by Isaac Slavitt where he tried both a brute force search and simulated annealing. Not surprisingly, I will try CPLEX on it. Here is the statement of the problem
A new city is being built which will include 20 distinct neighborhoods as shown... [More]
Tags: opl analytics programming modeling optimization 
Decision makers need decision support
How can you make an optimization application be accepted by decision makers? The answer I gave that in my last post was to provide interactive applications. It so happens that colleagues of mine already discussed that in an IBM book Optimization and Decision Support Design Guide I can't resist quoting some of it given how it captures what I tried to express in my previous post.
Decision makers need decision support
Decision makers will not use any analytics tool unless they trust it.... [More]
Tags: decision analytics optimization 
Interactive Optimization
In my last post I discussed how gamification could be used to overcome the resistance to automated decision making systems. The case discussed in my previous post was about a system that computes retail prices for hotel rooms. The point of gaming was to show that human intervention would degrade the business outcome. Prices set by humans lead to less revenue than prices set by the system.
Interesting comments on that post made me realize that I have been a bit extreme in my will to make a point. While I stand... [More]
Tags: decision optimization analytics 
We must show the pain before we can propose the cure
Part of my job is to inject optimization in IBM Anaytics solutions. During one of the discussions with solution teams we argued about a fairly general issue that can prevent prescriptive analytics adoption. I think it is worth sharing.
Specifically, one colleague presented the following analytics classification.
I said that we should rather use the one below (I discussed it in Prescriptive vs Predictive Analytics Explained .) where the question prescriptive analytics answers is " What... [More]
Tags: analytics optimization 
Optimization Is Ready For Big Data: Part 4, Veracity
Big Data promise is to enable better decisions based on data. The idea seems appealing yet there is a caveat: is the data reliable enough to base decisions on it? Question is to what extent can we trust data? My experience shows that cleaning data can take up to 80% of an analytics project. This is well known, and is often called the veracity dimension of Big Data . Point is that most data in the Big Data era is uncertain, see for instance the figure below, taken from a post by John Poppelaars... [More]
Tags: optimization analytics uncertainty big_data 
Optimization Is Ready For Big Data: Part 3, Variety
A colleague of mine once told me that Big Data should be called "All Data". Indeed, one of the key dimension of Big Data is to apply analytics techniques to all kind of data. Other dimensions include volume and velocity of data.
Can optimization be applied to all sorts of data? I'd say yes despite the fact that optimization primarily deals with numerical data. Indeed, optimization has already been applied to a wide variety of data, much more than common knowledge may suggest. Let's see... [More]
Tags: big_data optimization analytics 
Optimization Is Ready For Big Data: Part 2, Velocity
Proponents of Big Data boast about how it might help get personalized behavior from all the things and systems people interact with (web sites, mobile apps, customer support services, internet of things, etc) . These systems have to deal with data in motion such as web interaction, sensor feeds (eg body temperature), video, social media feeds, etc. Dealing with such data is the velocity dimension of Big Data . I have discussed how optimization could be applied to another big data dimension, namely large volume of data, in my... [More]
Tags: big_data prescriptive optimization analytics 
Optimization Is Ready For Big Data: Part 1, Volume
I had the honor to give a tutorial at a Big Data and Optimization seminar, thanks to an invitation from John Poppelaars , One of the topics I discussed seemed to resonate well. Let me try to explain it here.
The first thing people think of when they hear about Big Data is large data volume. There are other dimensions than volume in Big Data, see Big Data For Dummies for instance, but let's focus on large data sets. Is current optimization technology ready for the huge data sets available now... [More]
Tags: optimization predictive prescriptive analytics big_data 
Trying Decision Optimization on Cloud beta: part 3, OPL
We have been using CPLEX on the cloud in a previous post . Let us now look at models written with our OPL modeling language. Using a high level modeling language such as OPL can be more productive . It also enables the use of constraint programming if need be.
We assume from now on that we have both an IBM ID and a valid registration for IBM Decision Optimization on Cloud. If not, then please refer to the onboarding steps described here .
Let us go to the... [More]

Trying Decision Optimization on Cloud beta: part 2, CPLEX
In our previous entry we made a demo tour of our Decision Optimization on Cloud service. Let us now start real work with CPLEX. The first step is to get an IBM ID. If you don't have one yet, then you can start here . The next step is to go to the registration page that is accessible in various ways from the demo page .
It lets you create an IBM ID if you don't have one. Assuming you have got one, we move to the next page by clicking on the Already have an IBM... [More]
Tags: cloud analytics optimization oaas saas 
Trying Decision Optimization on Cloud beta: part 1, Demo
We just announced the open beta for our forthcoming Decision Optimization on Cloud offering. This is a major step forward in making optimization more consumable by born in the web applications. The first drop of our service enable operations research (OR) practitioners to solve their problems online via a very simple interaction. We will expose service APIs in future drops, stay tuned.
Let us look at how the service can be used right now. It assumes you have one or several optimization problems to solve,... [More]

2015 Prediction: Prescriptive Analytics Will Make It
This is prediction season. I never played that game so far, but felt compelled to do so after reading quite a few predictions about what will happen in 2015 around Analytics and Big Data. I won't repeat what seems to be a consensus, and will refer to two specific lists that I found more interesting than others.
The first list is by Nathan Brixius, from Microsoft. Here are the top items, I'll let you read Nathan's blog to get the meat behind the titles.
Adoption of higher productivity analytics... [More]
Tags: big_data prescriptive analytics optimization 
2014 Top Posts
The 5 most read entries for this blog in 2014 are the following.
The Analytics Maturity Model
Solving the hardest Sudoku  part 1
How Were Rosetta/Philae Operations Scheduled?
Memory Locality
Data Science Is Hard: A Look At Sotchi Olympics
You can access the top entries of all time via this link .
I would like to thank again all the readers that made... [More]

Analytics For Vehicle Routing (and for Santa)
How could we have helped Santa deliver its gifts as fast as possible? An OR practitioner will probably model Santa's problem as a traveling salesman problem ( TSP ). Indeed Santa needs to compute the shortest route through all kids homes. The same OR practitioner may also want to model that deliveries must occur at night. Instead of a TSP, the problem becomes a vehicle routing with time windows (VRPTW) problem. Although common wisdom, these models may miss the point as we shall see below.
Before discussing what... [More]
Tags: analytic prescriptive predictive optimization 
Prescriptive vs Predictive Analytics Explained
Imagine for one second that you are Loïc Peyron , the recent winner of the transatlantic yatch race La Route Du Rhum . What would you use to reach your destination as fast as possible? Of course, you would work on getting the fastest possible yacht. You would also train to maneuver your ship the best possible way. But once the race is on and you're on the sea, what tools would you use to decide in which direction you should steer the ship?
Credit: Wikipedia
... [More]
Tags: analytic optimization predictive prescriptive 
How Were Rosetta/Philae Operations Scheduled?
Unless you live unplugged you certainly saw the astounding pictures of Comet 67P/Churyumov–Gerasimenko taken by the Philae lander. Besides producing nice images, Philae embarked scientifc instruments , each developed by a European laboratory, to accomplish scientific experiments when approaching, and after landing on the comet. Given that communication takes about 25 minutes between Earth and Philae once landed, it was very important to carefully plan every operations of the mission in advance. Indeed, there... [More]
Tags: analytics constraint_programming 
Decision Optimization In The Cloud at INFORMS Annual Meeting
Update on Jan 10, 2015.
Decision Optimization on Cloud open beta is now available. I documented step by step on boarding here, but you can also try by yourself here .
Trying Decision Optimization on Cloud beta: part 1, Demo
Trying Decision Optimization on Cloud beta: part 2, CPLEX
Trying Decision Optimization on Cloud beta: part 3, OPL
Our INFORMS conference slides are available at our INFORMS 2014 Annual Meeting page
Original post:
The IBM Decision Optimization team will be... [More]
Tags: cloud cplex optimization analytics 
CPLEX 12.6.1 Announce
We are proud to announce a new release of CPLEX that will be available on December 5 for electronic delivery for paying customers. Date for academic initiative availability isn't known yet but we'll work to make it as soon as possible.
Official announce text is here . Enhancements include significant performance improvements across the board, as well as adding some licensing flexibility for our customers. In particular, enhancements include:
Improved performance of optimizers, notably MILP, convex... [More]
Tags: analytics optimization cplex 
Why Users Cannot Help You Improve Your Products
Making decision based on data seems a good idea, doesn't it? After all, this is the value promised by all Big Data promoters out there. Let's look at a real use case to understand better what might go right or wrong. I will focus on the decisions product managers must make when they think of the next version of their product. Should they base product evolutions on customer feedback?
Let's first address the case of disruptive technologies. It is (now) (well) known that the answer to the above... [More]
Tags: decision big_data sampling analytics design 
I Want The Best Solution
Who among optimization practitioners hasn't heard at least one customer saying "I want the best solution" ?
I certainly did.
When that happens one has to manage expectations because solving real problems to optimality may not be possible in a reasonable time. Customers must understand that finding the best solution and prove it is the best one is not always doable. They should instead focus on finding the best possible solution in a limited amount of time. If they agree to it, then they find it... [More]
Tags: optimization analytics moov 
Convex Optimization
I had the pleasure to be invited to attend the German OR conference last week in Aachen. There were many highlights at this conference beside the great venue and excellent organization led by Marco Lübbecke, see for instance Mike Trick on Laura McLay's semi plenary and Laura McLay on Mike Trick's plenary presentation . I will discuss here the semi plenary given by Stephen's Boyd on convex optimization.
Convex optimization is a generalization of linear... [More]
Tags: optimization theory 
Analytics Without Decisions
The view that analytic value comes from the decisions it enables is gaining momentum. I've discussed it in my Analytics Is A Mean To An End and The Analytics Maturity Model posts, but I was very pleased to read the following post from Seth Godin:
Analytics without action
Don't measure anything unless the data helps you make a better decision or change your actions.
If you're not prepared to change your diet or your workouts, don't get on... [More]

Machine Learning and Optimization
The view that Machine Learning is tightly linked to optimization is becoming common knowledge. There are training courses ( Stephen Boyd's and Steve Wright's ) mixing the two. The Data Smart book by John Foreman describes some key machine learning techniques as optimization problems using Microsoft Excel solver. This book is a great read by the way. But what made me write this blog entry is that I was quite delighted when I read the following from John Mount
My opinion is the best... [More]
Tags: machine_learning analytics optimization 
Price Optimization
Did you know that the price at which you buy your Coke bottle (pick any brand you'd like here) at your nearest retail store was probably set by a process that involved mathematical optimization? If not, then learn how it was probably done.
Let me first say that the idea of optimizing prices isn't really new. The airline industry has rolled out techniques called revenue management in the 90s, where the number of seats offered at a given rate was modified according to the demand. In a nutshell, if a given... [More]
Tags: analytics optimization commerce big_data 
Common Sense Operations
Is Optimization the tool of choice for improving operations? I tend to say yes, because we can be proud of the impact of optimization on various businesses and industries. See for instance the Franz Edelman Award finalists for examples of tremendous achievements . Yet, I regularly see nice alternative to using optimization techniques being successfully used. Let me give two examples.
I have been visiting a trucking company a while ago, pitching optimization as a way to significantly decrease mileage and... [More]
Tags: analytics optimization 
Memory Locality
How can a Java code be 85x slower than a C++ code solving the same problem? This post is trying to answer this question.
Why am I asking this question in the first place? It all started with a seemingly simple exercise. We were working on a large scale analytics (aka big data) project and had trouble agreeing on what results should a particular analysis return. I decided to write a C++ code for it, and a colleague decided to use Java. The goal was to use two completely independent implementations for cross... [More]
Tags: analytics graphs high_performance 
The Analytics Maturity Model
Analytics can be defined in many ways, but what matters is the purpose of analytics. Most definitions agree on the following: analytics is used to gain insights from data in order to make better decisions, see for instance INFORMS definition :
Analytics is defined as the scientific process of transforming data into insight for making better decisions.
Some speak of actionable insights to stress the purpose of such insights. Then, various levels of analytics maturity can be distinguished, depending on how much... [More]
Tags: optimization analytics 
Data Science Is Hard : A Look At Sotchi Olympics
Data Science is hard. I'll use an example that made lots of buzz to show some of the issues with data science. Two brothers, Dan and Tim Graettinger, who work for Discovery Corps, Inc. devised a predictive model that predicts medal count per country for the Sotchi Olympics. The Graettinger brothers model was commented on most data science and analytics sites, in OR blogs (see Laura McLay's entry) , even beyond . Question is: did they predict medal count correctly?
Before answering that question let me... [More]
Tags: data_science analytics 
Solving the hardest Sudoku  part 2
My previous post on Sudoku described how a fairly simple OPL model could be used to solve a hard Sudoku problem. I ended the post this way:
What?
What do you say?
I see, you're asking about the solution to the above Sudoku. Well, why not download CPLEX for free and run the above model to find out?
This post is a detailed tutorial on how to run that Sudoku model on a Windows PC using CPLEX. It also addresses an interesting challenge about using Microsoft Excel for defining the problem data... [More]

Solving the hardest Sudoku  part 1
Do you know the hardest Sudoku problem? Do you know the best way to solve it? Before answering these questions, let me remind you of what the Sudoku puzzle is about in case you haven't read a newspaper in the last decade (adapted from wikipedia ):
The objective is to fill a 9×9 grid with digits so that the digits in each column, each row, and each of the nine 3×3 subgrids that compose the grid (also called ""blocks") are pairwise different. The puzzle setter provides a partially completed grid,... [More]
Tags: constraint_programming mathematical_optimization optimization analytics mathematics sudoku 