by Soovin Yoon
After taking the Introduction lecture of this ISyE 823 class, I couldn’t help thinking about my first public sector OR problem. I was remembering that just as an ambiguous failure, but its details and reasons suddenly came to my mind due to class material.
When I was a sophomore, I took Scientific Management course that involves a term project to improve a bus service, named route K-02. As the university was located high up in the mountain, students had to take K-02 to reach the university from the subway station. There was two serious problems with the bus; it was always terribly packed, and it took a long and dangerous detour.
The cause of the detour was very clear. The route included a tiny apartment complex, named I-Apartments. I-Apartments was located on the hill, and the only path that connects it to the main street was very steep, narrow and winding. So students, faculties and staffs were complaining about their dangerous commuting route.
First it looked logical to simply remove the detour from the route to maximize the total satisfaction of the customer. Then people from I-Apartment will have to walk down a few minutes to use the bus, but since their number is quite smaller compared to university members, it won’t affect the average satisfaction that much. But it was unfair. K-02 was the only available public transportation for the resident of I Apartments to reach the subway station. The bus had a responsibility to provide the only service they can get.
The most popular solution from the class was installing a signaling system (a flag or traffic light), so that if there is anyone from I-Apartments waiting for K-02, then one can signal so that the bus can take the detour under the necessity. Otherwise the bus can skip the detour. This looked like a lovely flexible, reasonable solution. But thinking about how much additional attention the driver need to be in charge, how much unpredictable the bus schedule will be, and how upset a passenger will be if the passenger is late for the class and the bus suddenly takes unexpected detour – this was still unacceptable. It was too flexible.
One of the other solution was to remove the detour from K-02, and substitute its function of connecting the apartment to the subway by other route, K-04, which is less busy as it does not go to the university, and of which the original route is not very far from I-Apartments. This was also rejected as this involves too many changes in the system. Officials did not want to disturb the other bus route. It turned out not to be our decision variable at all.
In the end, we were unable to generate a solution good enough to be implemented. It was difficult for me at that time to understand why we had to fail. Now I am totally overwhelmed to see how even a simple and small problem like this can possess multiple issues of equity, determinacy and authority, as we are dealing with the public society.
Let me finish the post with short introduction about myself. I’m a PhD student in Industrial and Systems Engineering at UW. Now I am studying a dynamic priority reassignment problem in emergency medical services. And of course there are lots of issues other than objective maximization due to publicity of the problem. (For example, we don’t want to abandon a specific patient, just because abandonment actually helps us saving the resource and improving the objective function. We need to be there.) So here lies my expectation of the course. After taking the course, I want myself be a better decision maker in my next public sector operations research problem.
Soovin
Public Sector Operations Research 
I can understand how overwhelming that problem was – sometimes the bias of the status quo trumps good operations research modeling and analysis. I would argue with your conclusion (“In the end, we were unable to generate a solution good enough to be implemented”). Your reasoning and solution sounded like an improvement to me, but it’s true that it can be impossible to satisfy everyone with so many stakeholders and criteria that the current bad solution seems acceptable to all. Everyone was willing to risk death in a bus accident rather than make a small change! I’m glad you survived the dangerous commute to make it to UW 🙂
LikeLiked by 1 person
That was really hard problem, which you deal with even though it seems simple. It is good to have this implementation problem in the early stage in some sense for your future OR problems especially public sector ones. I believe you and your classmates learned to think more widely after you experienced this implementation problem in your early career. Unfortunately, we cannot put the human satisfaction or all stakeholders’ satisfaction as constraints in our models explicitly. But we are always trying to do our best while we also have other objectives that we need to achieve.. 🙂
LikeLiked by 1 person
At first glance, this problem has a few reasonable easy solutions…I personally would advocate requiring the residents of I-Apartments to walk a short ways to the bus stop. However, it’s virtually impossible to come up with a solution that pleases everyone. The apartment residents don’t want to walk. The officials don’t want to change other bus routes. The bus drivers don’t want to pay attention (that sure does sound weird).
Solutions to problems like this, often justifiable from an ORMS standpoint, are often dismissed by a refusal to accept change, not unlike the New York City-RAND Institute’s experience with the NYPD (Green and Kolesar, 2004). Individual satisfaction really should be sacrificed for a disproportionate increase in general satisfaction. It sounds like you did everything you could.
At any rate, I’m sure a lot of the issues with the current system can be attributed to the initial implementation of a less-than-optimal bus schedule.
LikeLike
Your comment gave me a better connection between the class material and my experience. While going through readings for the first week, I felt quite confusing to observe that relatively few of the studies had direct influence on policy-making in practice, although many models are so successfully done. That’s still compelling but less confusing now, thinking how difficult or impossible it is to satisfy all stakeholders.
About the solution to ask residents walk few minutes(As I didn’t elaborate it enough in the post); that was actually attractive, numerically. But after the field investigation, we observed that a decent number of the residents using bus were the elderly, so we felt somewhat.. awkward to ask them to walk up and down the steep path to reach the bus.
LikeLike
As someone with over 10 years working in the public sector, reading your description of your experience made me want to stand up and applaud your efforts, as well as your careful, deliberate approach to the problem. What you’ve described encapsulates the work of policy analysis and change management. The good news is, sometimes improvements DO get implemented. Also, in my experience, improvement initiatives tend to be cyclical, so even if a process or policy improvement isn’t implemented right away, sooner or later a new set of decision-makers will ask the question “hey, can’t we do something to make this better?”, at which point you dust off your past findings and look like a rock star.
I agree with previous comments that your work was not the issue, but rather that even the best ideas sometimes fail to gain traction right away. It’s the periodic wins that keep us going through the setbacks, and I’m sure you’ll have many wins in the future. Keep up the good work.
LikeLike
Thank you so much for making a comment that’s so cheerful! I have a great hope that there will be chances to observe wins – my improvements being implemented – in the nearest future.
LikeLike