Punk Rock Operations Research T-shirt

I enjoyed designing and wearing a Punk Rock Operations Research T-shirt at the INFORMS Annual Meeting. I brought a few extras and sold out at the conference. If you want to purchase a T-shirt, you can order one here at CafePress.


Ed Kaplan, INFORMS Member-In-Chief

I am honored that Ed Kaplan, INFORMS Member-In-Chief, wore my T-shirt!


Matt Saltzman purchased a T-shirt but wasn't wearing it yet in this picture.

Matt Saltzman purchased a T-shirt but wasn’t wearing it yet in this picture.

10 things you can do with an industrial engineering degree

Kim Christopher (BE IE, MBA) won my department’s Distinguished Achievement Award for alumni. She visited my department and gave a seminar about her career. She had worked in many roles in many industries and talked about her experience.

Kim’s list of what you can do with an ISYE degree:

  1. You could manufacture products.
    • Kim interned at General Motor, Procter & Gamble, and the military (she made motors for the F14 fighter jets)
  2. You could research new technologies.
    • Robotics, expert systems, intelligent transportation systems.
  3. You could sell products or services.
  4. You could market or communicate technical things.
    • Engineers companies need help telling their stories.
  5. You could develop new products.
  6. You could improve quality.
  7. You could run a business.
  8. You could teach.
  9. You could have a dual-career family.
    • I loved Kim’s enthusiasm for wanting it all, having it all, and not apologizing for it.
  10. You can give back.

Other takeaways:

  • Some of Kim’s achievements were due to being opportunistic when good opportunities came by. She was able to implement an intelligent bus transportation for public transit in Napa Valley because only public leaders there were receptive.
  • Sales was more fun than engineering.
  • Kim worked on an MBA while working full-time. It was grueling and expensive, and she has no regrets.
  • Kim said her choice of a husband was her most important choice along the way.

Kim Christopher

Kim Christopher

analytics for governance and justice

In May 2016, the Office of the President released a report entitled “Big Data: A Report on Algorithmic Systems, Opportunity, and Civil Rights” that challenges the idea that data and algorithms are objective and fair. The report outlines President Obama’s plan for identifying and remedying discrimination in data and automated decisions by making data and processes more open and transparent.

This is part of the White House’s plan for data driven governance. With better data, you can make better decisions. I love it.

President Obama said that “information maintained by the Federal Government is a national asset.” He started data.gov, which is a gateway to government agency data to researchers and the public.

Created as part of the President’s commitment to democratizing information, Data.gov makes economic, healthcare, environmental, and other government information available on a single website, allowing the public to access raw data and use it in innovative ways.

Data.gov began as a tool to reduce government waste, but it has since branched out to meet other goals, such as the aforementioned social justice issue inequities. The White House created the position “Chief Data Scientist” and hired DJ Patil to fill the position.  He has been working on breakthroughs for cancer treatment lately.  The White House hosted an “Open Data Innovation Summit” in September 2016 to share best practices regarding the opening up of government data. While I applaud the trend of open data, it is necessary but not sufficient for reducing inequities, informing decisions, and cutting government waste.

I am less familiar with the big wins that data driven governance has had. Please let me know what they are in the comments. I have no doubt that there are big wins. With better data, we can make better informed decisions.

Data is a huge topic, and there is a lot of data out there. The government investing in archiving and analyzing data is necessary for breakthroughs to happen. There are a lot of people involved in this effort. My colleague, Dr. Patti Brennan now heads the National Library of Medicine. The National Library of Medicine is composed of data to support medical research, and I’m glad we have a Wisconsin ISYE Professor Emeritus and rockstar in charge.

I started this post before the election. I hope the project continues its momentum in the next administration to have an impact. Only time will tell.



Data topics at data.gov

Data topics at data.gov

final Presidential election forecast predictions

The Presidential election forecasting models I’ve been following this election cycle are all pointing toward a Clinton victory. Now we have to wait and see.


Election Analytics @ Illinois

Princeton Election Consortium (Sam Wang)

FiveThirtyEight (Nate Silver)

New York Times Upshot forecast

Daily Kos (Drew Linzer)

David Rothschild’s prediction market forecasting model

Huffington Post Election Forecast

Sabato’s Crystal Ball

13 Keys to the White House

Why don’t all of these models agree? A few articles I’ve read lately about forecasting models and polling:

The Oregon Trail Knapsack Problem

In 1999, Jennifer Burd, John Ainsworth, Brian Casto and Sheau-Dong Lang published a paper entitled “Experiments with the ‘Oregon Trail Knapsack Problem’ ” Unfortunately, the pdf of the paper is low quality, so the paper is difficult to read, but it’s a fun election distraction.

There will be blood.


Abstract. This paper presents hybrid algorithms for a variation of the Bounded Knapsack Problem which we call the Oregon Trail Knapsack. Our problem entails imposing a cost as well as a weight limit, constraining the values of types of items by means of a variety of value functions, and allowing the value of one item type to be dependent on the presence or absence of another type in the knapsack. These modifications to the original problem make it more complex and require adaptations of known knapsack algorithms. To solve this problem, we combine constraint propagation techniques and domain pruning with classic branch and bound approaches that require a sorting of the items. Our experiments compare a constraint-language implementation with a simulation of the constraint-based system in a procedural language. Results indicate that the constraint-based solution is natural to the problem and efficient enough to solve large problem instances typical of the application.

Check out the iOS game or play online.

I vividly remember playing the game in grade school and recall that the game was stochastic, particularly if you needed to hunt. In addition, you had to make sequential decisions regarding hunting: do you extend your trip by a day so that you can eat? At some point, you had to hunt or you faced starvation. I was bad at hunting. I usually chose the robust strategy of playing a merchant so that I could purchase my food when I needed to. I did not starve, but I also did not achieve as many points as I could if I was not playing a wealthy character. From what I recall, no one escaped dysentery.

What was your Oregon Trail strategy?


how to suppress the vote through bad resource allocation

Suppressing the vote is in the news a lot this election. Several media organizations have reported that the Trump campaign is trying to suppress the vote in an attempt to win the election. Voter suppression can occur through discouraging people to vote by running a negative campaign or disenfranchising voters by claiming that the system is rigged and votes don’t really. Votes can also be suppressed by managing resources in a sinister way.

There were very long voting lines in Ohio (a battleground state) in 2004, 2008, and 2012 that led some to question state officials motives. A 2008 NY Times article by Adam Cohen outlines some of the allocation issues in Ohio on Election Day in 2004. One way to suppress the vote is to under-allocate voting machines to precincts, which leads to voting queues exploding. As a result, many voters will balk or renege before voting, thus suppressing the vote. Under-allocating voting machines in a targeted set of precincts can suppress the vote of one political party.

Bad resource allocation can suppress the vote, whether it is intentional or not. And that’s a problem. Good planning doesn’t just to efficiency and effectiveness; it also leads to equity — everyone having equal opportunity to vote and equal access to a relatively short voting queue in their precinct.

Queue basics: the voters are customers who enter the system, where the system is a precinct’s voting location with several voting machines or booths. The voters wait in a queue to cast their votes in voting booths (the servers). The queue can get long due to one or more of these three factors:

  1. The arrival rate can be too high
  2. Voters take too long to cast their vote
  3. There are too few voting booths

We generally assume that the arrival rate and the voting time are exogenously given input parameters and that the only thing we can control is #3, the number of voting booths at each precinct. This is mostly true. Technically, we don’t have much control over factors #1 and #2. Voters show up to vote when they want and take the time they need to cast a vote.  However, a precinct with a very long line can encourage voters to balk (not enter the queue) or renege (leave the queue after waiting for a bit), which means that we have some control over the effective voting rate. We can lower the effective voting rate by allocating too few voting booths to a precinct. This is how you can suppress the vote.

You can also suppress the vote by not anticipating a longer ballot, since the time it takes to vote is largely determined by the length of the ballot. A long ballot or a ballot with referendums that are worded in confusing legalese can slow down the voting process and lead to very long queues. It’s important to study the ballot length and request more voting booths if the ballot is long.

A 2013 paper by Muer Yang, Michael Fry, David Kelton, and Ted Allen called “Improving Voting Systems through Service‐Operations Management” studies how to allocate voting machines to precincts using queueing models and simulation models. All authors were at Ohio universities at the time, and they develop a method for allocating booths to precincts that are robust to assumptions regarding the potential arrival patterns of the voters and the possibility of voting machine failure.

One way to avoid waiting in line on Election Day is to vote early or with an absentee ballot. I hate waiting in line so I already did this. NPR has full coverage of how to vote early. Many–but not all–states allow for early in-person or by mail voting with no excuse necessary. Voting early helps to reduce the arrival rate on Election Day, which reduces waiting times while not suppressing the vote.

Yang M, Fry MJ, Kelton WD, Allen TT. Improving Voting Systems through Service‐Operations Management. Production and Operations Management. 2014 Jul 1;23(7):1083-97.


Related posts:

Why you see Christmas items in stores during the summer: a story about supply chains

I confess that I don’t mind Christmas items in stores before Thanksgiving. While I agree that it’s too early, I understand why it’s sometimes a good business decision. Christmas items are similar to other “perishable” items like newspapers, produce, blood, and fashion, that go bad after Christmas in this case. These supply chains are managed differently than the supply chains for non-perishable items. Anna Nagurney has some excellent posts on perishable supply chains, blood banks, and blood supply chains.

I discovered an article about this by Kori Rumore in the Chicago Tribune that goes into detail about Christmas supply chains. Below I list a few supply chain observations I thought about or learned when reading the article. I use “Christmas” here generically because “holiday” is ambiguous, although other winter holiday items are often included with Christmas items (Christmas items make up the bulk).

  1. Christmas and holiday items are seasonal items in stores. Many box stores have a fairly large section dedicated to only seasonal items, where they put out many items for the next season. A typical seasonal section may rotate through the following merchandise: Valentine’s Day, Easter/spring, summer/beach, back to school, Halloween, and Christmas items. After Halloween, the next “season” associated with a lot of merchandise is Christmas. It’s better to stock the seasonal section with Christmas items than let it remain empty.
  2. Some stores like craft stores also have to deal with the crafting supply chain: you have to make holiday decorations well in advance of the holiday, so the holiday craft supplies need to be available for purchase crazy early. Christmas craft supplies often come out in the summer. I’m a regular at the craft store, but this never fails to surprise me.
  3. Seasonal items can sell out, so many consumers like to shop early while there a selection of items, and retailers have an incentive to cater to early shoppers (even if this irks most shoppers).
  4. The holiday items you see in the summer may be a tiny fraction of what you will eventually see after Thanksgiving, when the bulk of Christmas items grace store shelves. Some retailers only put out a few select items, such as collector’s items. The US Postal service and Hallmark sell to stamp and ornament collectors, respectively, and they put out Christmas items early for the collectors, not non-collectors like me. It had never occurred to me that I might not be the target audience for those early Christmas items.

The Chicago Tribune notes that 40% of consumers like to start holiday shopping before Halloween. I am one of them, but I usually shop for gifts, not holiday items. I’m a bargain hunter and usually buy holiday items like cards and wrapping paper when they are deeply discounted (the day after Christmas).

On a final note, supply chains are not why stores sometimes play Christmas and holiday music before Thanksgiving. There is no excuse for that🙂

Where have you seen many Christmas items for sale?