For the Naive Algorithm submit a transcript showing the output from your program (the closest pair of points, the distance, and the running time in milliseconds) for inputs of size: 10, 20, 50, 100, 250, 750, 1500, 3000 and 6000.
For the Divide-and-Conquer algorithm if you complete the whole thing (worth 70% of the grade) then do exactly as stated above for the Naive Algorithm.
If you completed the Splitting Part of the Divide-and-Conquer algorithm (but not the rest) then for inputs of size 10 and 20 output the original input (sorted by x-coordinate) and then output the points in both the left and right halves sorted by both X and Y coordinates. (XL,YL,XR,YR in the notation of the lab). Notice, that you'll just do this for the original input. So when n=10 there will be 5 points per half. When n=20 there will be 10 points per half. Once you have this working correctly, then you'll be ready to make the recursive calls and then combine. The idea of separting this part is so that you first debug the dividing part before going further.
For The Comparison you should submit a graph (easily created via Excel or the program of your choice) showing the running time in milliseconds of both the naive and divide-and-conquer algorithms as a function of the input size and the minimum, maximum and average value of the running time of the divide-and-conquer algorithm over 100 randomly generated inputs of size 1000. (Just put a loop in main to repeat the experiment 100 times. You'll get a different input each time as long as you don't seed the random number generator in the loop.) Write a sentence of two about how imprecise the graph is by not averaging over many inputs for every size tested.
NOTE: In creating the graph, the assignment tells you to have n go from 25 to 6525 in intervals of 250. These values were selected so that if students run this on a CEC shared machine (e.g. Hilton, Clarion, ...) that it wouldn't use too many cycles. If you are running this on your own machine and would like to go to a value bigger than 6525 feel free to modify this range. For example you could go from 25 to 26025 in intervals of 1000. If you want to go even higher, go ahead.
Expore the Crossover: Write a brief discussion about what crossover you found, give the portions of code that you changed for your "hybrid" implementation and report the results coparing the "hyprid" implementation with the divide-and-conquer approach. Showing a graph would be a good way to do this.
Change the Crossover: Describe your modification to speed up the naive implementation, the code for your modified naive algorithm, and a discussion of whether/how much the crossover point moved.

Return to the CS 241 FAQ Page