16 tablets were deployed in class, with two students at most tablets (there were a few groups of three, and a few singletons). There were no technological problems - everything went smoothly. In general, participation was very high. There was also significant collaboration on the problems. The weakness that I felt on this lesson is that I did not take advantage of their solutions - the activites all had a generally correct answer - so displaying them did not have much value.
This is the activity that was used in the five problems lectures delivered at Beihang. The goal of the activity is to verify that students have understood the definition of the problem. I hope that almost all students succeed in the activity.
After class notes:
Students uniformly got this one correct - it was easy, and this
established that students understood the problem.
The problem is to simulate greedy algorithms with different heuristics. This activity is to show that a collection of different heuristics do not work for finding an optimal solution.
After class notes:
I set up this activity by first generating a small example by hand to
show the process of simulating an algorithm, then I continued talking
about the activity after I put the slide up. This activity was more
challenging than the previous one, although all students got the
correct solution. Algorithmic simulation actually works quite well as
an activity. More direction could have been given to have students
indicate when the different tasks were removed.
This activity is to demonstrate understanding of the problem, and possibly to get insight into the eventual algorithm. The problem is big enough, that students may start using an algorithm to compute the labelling. I am not sure if I will tell the students verbally what the answer is in advance.
After class notes:
This student was phrased as "can you find a solution of size five".
The example was complicated enough that it was hard to get the answer
right without using the proper algorithm. A number of students seemed
to use the proper algorithm - this could have been either discovery,
or from the text book. It was difficult to evaluate the submissions.
The use of colors (and especially the highlighter) was key.
Although this is a proof problem, it is pretty easy. The expected solution is just to observe that three overlapping jobs cannot be scheduled with two processors.
After class notes:
This gets the record for the fastest student submission exercise. Work and display time were estimated to be about one minute by Valentin. I thought this was quite effective in getting many students to express the lower bound idea. The proof idea was just to identify an obstruction that would require at least three processors.
I do not anticipate reaching this one - just in case I go much faster than anticipated, I have an example ready for students to explore the current problem.
After class notes: As expected, this was not reached. Timing worked well so I could just introduce the overall problem, without getting to the technical details.
There were a few additional submissions before and after class, as well as a few creative submissions during activities: Student submission examples