how queueing theory helped me with work-life balance

I find that mathematical models used to solve systems engineering problems are also useful for managing personal issues I face in everyday life. Stochastic processes and queueing theory helped me find and maintain a healthy work-life balance.

I was often wracked with guilt for taking personal time during graduate school and on the tenure track. Three things really helped.

1. Role models who are successful and who have balance made a big impact. I had an advisor in graduate school who talked about how I should take time off every week, and that was a help even though I often did not follow his advice. I am constantly inspired by colleagues I know who are OR rockstars and also have a healthy home life and engaging hobbies.

2. Having children helped me set healthy limits at work at a young age. I had my first child in graduate school, and in retrospect, that was probably the best thing for helping me find balance personally and professionally. My daughter needed me, and I needed to be there for her. I know the statistics indicate that maybe I am an anomaly, but I have had to balance the demands of work and motherhood throughout my professional life. Learning this early helped me later on.

Note: I am not suggesting that if you are really struggling with work-life balance issues you should have a child. I am certain I was predisposed to handle the transition to parenthood 🙂

3. Caretaking is not the same as taking time off from work; it’s a different kind of work. This is where stochastic processes and queueing theory helped.

I am a single server queue who works day after day. Work enters my queue with rate \lambda and I finish tasks with rate \mu . If the arrival rate of new work exceeds my service rate (i.e., if \lambda / \mu > 1 ), the queue explodes and eventually becomes infinitely long. I can assure you that \lambda / \mu > 1 .

One (unhealthy) way to frame this situation is to note that one way to reduce the length of the queue is to work all the time. But this is futile in the long run.

Another way to frame this situation is to note that I have some inefficiencies in how I work. Over time, I have become more efficient and have increased by \mu .

The third–and best–way to frame this is to accept that at some point, I can no longer improve my efficiency. And at that point, it’s still true that \lambda / \mu > 1 . Queueing theory demonstrates that there is no systematic way to “catch up,” so I might as well take a break to run or go to the symphony or make cookies.

Having said that, I still struggle with guilt if I take breaks longer than one hour. I’m a work in progress 🙂

What helps you find balance?

 

Advertisements

Pre-tenure planning for your post-tenure life: my interview for the Decision Analysis Society

Allison Coffey Reilly and Florian Federspiel interviewed me for the INFORMS Decision Analysis Society’s quarterly newsletter as part of an article about how OR faculty transitioned from pre-tenure to post-tenure life. My interview is below.

~

Pre-tenure planning for your post-tenure life

For this edition of Ask DAS, we had the exciting opportunity to speak with Drs. Laura Albert (University of Wisconsin – Madison), Jason Merrick (Virginia Commonwealth University, Canan Ulu (Georgetown University), and Jun Zhuang (University at Buffalo) about how they thought about their transition from pre- to post-tenure. We are sharing an excerpt of our conversation below. The transcript has been edited for brevity and clarity.

  • Was there anything that surprised you while preparing your tenure dossier or when going through the tenure process?

Laura Albert (LA): I went through the tenure process twice in two years – first at Virginia Commonwealth University and then at University of Wisconsin – Madison, after I moved. This is not the typical way to do it. The process was significantly more difficult the second time around. If you’ve been at an institution for five years, the process should be fairly straightforward. The institution, college, and department should provide clear expectations during that time. They might provide example portfolios – nothing should be too surprising. But, when a faculty member switches universities mid-stream, that can pose unique challenges. He or she may not be aware of or be able to meet specific institutional or departmental milestones at the new universities. These need to be made clear and then discussed and managed by the new university.

  • It is frequently said that tenure is needed to allow for researchers to take-on riskier research. In what ways did your research or your approaches to research become “higher risk?”

LA: Yes, definitely. Post-tenure is a good time to take on new research projects. One of the difficulties with new research is that there is high start-up cost – it might take 6 – 12 months to scope out the work and get things going. It’s a much better time to get more irons in the fire, which can be very intellectually rewarding. The research itself might not be risky, but the time it takes to get the research going could be seen as a risk if done before tenure.

  • Were there research topics that interested you pre-tenure that you knew you shouldn’t have broached until post-tenure? How did you conclude pre-tenure to wait to pursue the topic?

LA: I have a foot in bracketology, including football rankings, basketball rankings, college football playoff forecasting, and that was totally a post-tenure gift to myself. It such a fun thing to do, mainly for outreach. It gets people excited for industrial engineering and shows people what we do in operations research. Related to that, I developed a course on sports analytics. It’s good to develop some things post-tenure that might not be directly related to your research or that you, per se, having funding to support, but that really brings you enjoyment. For me, it has helped me to establish a balance that I like in my life.

  • Do you think it’s necessary to have a plan prepared pre-tenure for the post-tenure life? If so, how did or would you go about that planning?

LA: I guess, in some ways, I went from having a plan to being more flexible, which has been fun. I’m someone that really focused my time during my pre-tenure period. I had a lot of ideas, but I really only focused on one to two. I wanted to open that up post-tenure. I have been able to be more flexible in the ideas that I pursue. It certainly poses some risk – it doesn’t guarantee that you will get the maximum number of publications out next year that you could otherwise.

  • What do you think is the most misunderstood aspect of post-tenure life?

LA: Oh, you think you are going to be less busy. You’re busier than ever. All of those things that you do to get tenure make people take notice. So, they ask you to do more reviews, serve on more committees, and your department will expect you to do more. And there are often a bunch of research ideas that you have been waiting to pursue.

  • Was there anything that I missed?

LA: I had three children pre-tenure. I did get nervous any time someone brought up the conversion of tenure or children, because children are so often framed as having a negative impact on tenure. There are a lot of exceptions to what we think is the rule. It’s certainly possible to have children on the tenure clock and succeed.

 


how long will volleyball games last with side out vs. rally scoring: a Markov chain approach

I introduced a Markov chain to model volleyball scoring schemes in my course on probability models. I am old enough to remember side out scoring, where a team could score a point only if they served the ball. If the team did not score the point, there would be a side out, and the other team would serve the ball in recognition of winning the play. Games were played to 15 points. Under rally scoring, a team scores a point for every play they win regardless of who served. Games are generally played to 25 points. USA volleyball switched from side out scoring to rally scoring in 1999, and college and high school volleyball switched soon thereafter. More here.

Years after the switch, I met Phil Pfeifer from Darden. We discovered that we both played a lot of volleyball in graduate school. There was a push to rally scoring (then called Fin-30) when Phil was in graduate school well before the switch. He wrote a paper* in 1981 that modeled the two scoring schemes as a Markov chain to identify the situations under which rally scoring would leader to shorter or longer games. Because rec sports could be so lop-sided, it was hypothesized by some that rally scoring would lead to longer games. This is counter-intuitive. The Markov chain analysis confirms that this is true in some circumstances. The games tend to be longer under rally scoring when the serving team tends to hold the serve or when the teams are lopsided.

My slides from class are below. In my slides, I also examine the benefit of serving first under side out scoring. Since the serving team is the only team that can score points, the team that serves first is one step closer to winning. That is an enormous advantage if there is a high probability that the team that starts with the ball holds the serve.

* P. E. Pfeifer and S. J. Deutsch, 1981. “A probabilistic model for evaluation of volleyball scoring systems, Research Quarterly for Exercise and Sport 52(3), 330 – 338.


vote early, vote often

Legally, you can only vote once. But if you vote early, you can enable more than one vote to be cast.

Voting in election day is an application of queueing theory. When voter turnout is high, as it is expected to be this year, the queues can become long. Sometimes very long. The lines in Ohio in the 2004 election are infamous. As a result, many voters balked* or reneged** before casting their votes. Alexander S. Belenky and Richard C. Larson ask: Did election queues decide the 2000 and 2004 U.S. presidential elections? Their analysis is summarized in their ORMS Today article.

Practically speaking, queue lines can be reduced one of three ways:

  1. Fewer voters enter the queue
  2. There are more voting booths and people processing voters
  3. Ballots are shorter

Voting early or voting absentee shortens the queues on election day by addressing issue #1. So while you can cast only one vote, casting your vote early means that you can keep the queues shorter on election day and possibly enable someone else to vote who otherwise would not be able to. This is meaningful in practice, since many voters cannot wait in line because of family responsibilities or shift work. So far, 2018 seems to be setting records for early voting. I voted absentee because I will be in Pheonix for the INFORMS Annual Meeting on Election Day.

 

* Balking: The voter decides not to enter the waiting line.

** Reneging: The voter enters the line but decides to leave before voting.

 

Related posts:


Annie Duke on playing poker and making good decisions under uncertainty

Poker player Annie Duke appeared on the Slate Money podcast this week to discuss decision science, and it reminded me that I should blog about her. On the podcast, she talked about how thinking in terms of betting can help make everyday decisions and decisions in engineering systems.

I first heard of Annie Duke on The Moth podcast, where she talked about imposter syndrome as one of the few women who played poker professionally. Her story on imposter syndrome was compelling, because in a tournament on television, her cards were shown on camera during play. This transparency made her vulnerable to judgment, because the viewers could assess her choices with full information and possibly second guess her strategies. She didn’t just feel transparent, she was transparent.

Annie Duke pursued a Ph.D. in cognitive science, which she ultimately did not finish, and has written about how to make decisions. Her knowledge of decision-making helped her to be one of the best poker players ever. Duke uses heuristics instead of optimal decision strategies, which is appropriate for playing poker and making decisions in real-time. She understands that humans are not rational decision makers who think differently about short term decisions versus long term decisions. She wrote a book called “Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts” about decision making under uncertainty.

In many of her interviews, such as her interview in Slate Money and in the video below, Duke articulates how the decision-making strategies she used to win at poker can inform decisions in systems, including engineering systems. In poker, quite often having a good process does not guarantee good outcomes. In poker, unlike chess, information is incomplete and other plays may be adopting sub-optimal or deceptive strategies.

For more reading: Decisions sciences was instrumental in substantially improving predictions in the Netflix prize.

 


Punk Rock OR goes to Heinz College

I visited the Heinz College at Carnegie Mellon University to give a talk about my research on cyber security and trustworthy computing.

My talk was entitled “Models and algorithms for protecting critical information technology infrastructure”

Abstract
This talk is motivated by a cyber-security planning application, where we explore how to mitigate vulnerabilities within information technology (IT) supply chains for securing cyber-infrastructure. To do so, we formulate new optimization models based on the coverage models and network interdiction models. In this research, we investigate how to identify a best combination of cost-effective mitigations that maximally delays supply chain attacks when there exist multiple adversaries. We present new Stackelberg game models that explicitly formulate the interaction between a defender and multiple attackers. We propose max-min interdiction models for critical infrastructure protection that prioritizes cost-effective security mitigations to maximally delay adversarial attacks. We consider attacks originating from multiple adversaries, each of which aims to find a “critical path” through the attack surface to complete the corresponding attack as soon as possible. Decision makers can deploy mitigations to delay attack exploits, however, mitigation effectiveness is sometimes uncertain. We propose a Lagrangian heuristic that identifies near-optimal solutions efficiently.

I discussed the following two papers in my talk:

  1. Zheng, K., Albert, L.A., Luedtke, J.R., Towle, E. 2017. A budgeted maximum multiple coverage model for cybersecurity planning and management.
  2. Zheng, K., and Albert, L.A. 2018. Interdiction models for delaying adversarial attacks against critical information technology infrastructure.

I had a delightful visit. I have visited Pittsburgh several times before, and I always enjoy seeing the Cathedral of Learning. The highlight this time was meeting with the faculty and students at CMU. Dr. Alex Jacquillat was my faculty host. Carnegie Mellon is a university with a lot of collaboration, and this was evident during my visit. My schedule included meetings with faculty and students from Heinz College, the Tepper School, computer science, and engineering.

I saw Dr. Al Blumstein of Heinz College give a talk about criminal justice and operations research when I was a graduate student, and he is part of the reason why I pursued research in public safety in emergency medical services. I gave my seminar in the Alfred Blumstein Classroom at CMU. It was an honor.


Punk Rock OR is on PBS!

I do not believe teaching is confined by the walls of the classroom or the boundaries of the university. I am passionate about talking to the public about science, engineering, operations research, and analytics. I especially enjoy talking about my research. Operations researchers and industrial engineers make important contributions to basic science and important applications, yet what we do has been a well kept secret. This is why I agree to give public talks about my research whenever I can. Doing so gives me an opportunity to educate the general public and improve scientific literacy. I have always hypothesized that most people would appreciate the work we do if they knew more about it. My experiences suggest that my hypothesis is true.

I gave a talk that was recorded on campus in the Spring. It aired on the University Place series on July 23, 2018 on Wisconsin Public Television, Madison’s Public Broadcasting Service (PBS) station. My talk aired the same day as the Great Wisconsin Baking Challenge: Week 5 Pies Recap. What great company! You’ve gotta love public television 🙂

Here is my talk title and description:

Advanced Analytics: From Emergency Response to Brackets

University of Wisconsin-Madison Industrial and Systems Engineering professor Laura Albert will talk about how engineers use math models and analytics to solve problems and design systems. She will provide an overview her discipline of operations research and advanced analytics and will discuss its wide ranging applications, focusing on examples from her research that addresses problems in emergency response and bracketology.

You can watch the video on PBS using the link below. The talk has closed captioning. My dad endorsed the video and told me he particularly liked the graphics in my talk.
https://player.pbs.org/viralplayer/3014502909/

The unedited talk is on YouTube:

I gave an earlier version of this talk at the Middleton Public Library for their “Scholar’d for Life” lecture series. The lecture series is in partnership with the UW Madison Speaker’s Bureau (my profile is here). Taking the “Wisconsin Idea” as its starting point, this series aims to promote lifelong learning, intellectual curiosity, and engagement between academics and the community as a whole.

The organizers at the library asked me to speak about applications of industrial engineering. I thought that no one would show up for a talk marketed like that. I was wrong. About 100 people showed up for my talk. It was a packed house! The attendees were really engaged and asked me many questions after the talk.

I was thrilled that a few kids in middle school and high school came to my Middleton Public Library talk, including girls. I try to embody the spirit of the #ILookLikeAnEngineer movement in my public engagements to challenge stereotypes about engineering. One of the girls who attended the talk told me I reminded her of Lucy Wilde from Despicable Me 2 and showed me this picture:

I’m not sure if Lucy is my doppelganger, but I certainly like her hair and style 🙂