• Be able to apply asymptotic time complexity analysis to choose among competing algorithms.
• Be familiar with engineering applications for many of the fundamental algorithms discussed in the course.

This lab assignment is designed to stimulate creative solutions to problem that is difficult to solve efficiently.

While asymptotic analysis allows us to evaluate an algorithm's efficiency, it ignores other important issues like implementation complexity and time constants. In this assignment, you should focus on developing an algorithm that will produce the correct answer, scale well to large data sets, and run as quickly as possible on the given data set (listed in order of priority).

The GPS data available to you consists of a file with multiple trips. Each trip represents one commute from my home to MSOE or back. The first few lines of the file are:

Start time: 2004-03-05 11:09:23
record,fix,hour,min,sec,msec,latitude,longitude,alt
1,3,11,9,24,710,43.0835367,-88.0390017,702.09
2,3,11,9,26,310,43.0835367,-88.0390017,702.09
3,3,11,9,27,310,43.0835367,-88.0390017,702.09
...

The first line indicates the starting time for the first trip. The second line contains labels for the different fields associated with each record. The third line contains the data associated with the first data record. Each record includes (separated by commas) the record number, the number of satellites used to get a fix on the GPS receiver's position, the time the record was made (in hours, minutes, seconds, and milliseconds), the latitude, longitude, and altitude. These records continue until the end of the trip. Data for the second trip follow immediately after the first trip. In this particular example, the first trip has 1336 records. Here are lines 1337-1345 of the data file:

1335,4,11,31,52,750,43.0445550,-87.9081050,639.76
1336,4,11,31,53,750,43.0445550,-87.9081033,639.76
Start time: 2004-03-05 17:10:35
record,fix,hour,min,sec,msec,latitude,longitude,alt
1,3,17,10,36,380,43.0445983,-87.9081033,583.98
2,3,17,10,37,380,43.0445950,-87.9081017,583.98
3,3,17,10,38,380,43.0445917,-87.9081000,583.98
4,3,17,10,39,380,43.0445917,-87.9081033,587.27
5,3,17,10,40,380,43.0445850,-87.9081067,587.27

I would suggest that you begin testing with the smallgps.zip file which contains a few very short trips. At a minimum, your program should be able to produce all of the required statistics for this file. This file contains a subset of the biggps.zip and may be use for testing.

• A trip is defined as a collection of GPS positions tracking the position of my vehicle from when the car is started until it is turned off. Each trip in the data file begins with Start time: ....
• Two trips are said to be on the same commute if and only if both trips begin at the same position and end at the same position.
• Two trips are said to be on the same route if and only if every position in the first trip is also in the second trip and every position in the second trip is also in the first trip.
• Generally speaking, the data file consists of two commutes (from my home to MSOE or from MSOE to my home). However, the data file may contain some additional commutes (e.g., from MSOE to church). The most common commute is defined as commute with the most number of trips in the data file. If multiple commutes tie for the most popular commute, one of these commutes may be selected as the most popular.
• The most common route is defined as the route with the number of trips in the data file.
• GPS positions should be consider equivalent if they are within X meters of each other. Specifically, two trips are from the same commute if and only if their starting positions are within X meters of each other and their ending positions are within X meters of each other.
• Your program should be able to work with arbitrary GPS trip data provided it conforms to the format described above.

Your program should read a data file with GPS trip information (like the input file above) and indicate the following:

• Number of unique starting positions.
• Number of unique ending positions.
• Number of unique commutes.
• For the most common commute:
  • Number of unique trips.
  • Starting position.
  • Ending position.
  • Number of unique routes.
  • Number of unique trips for the most common route.

You should calculate the starting/ending position as the geometric centroid of all the positions considered to be the same position (using the X meter rule).

You should submit a partially completed lab report (fill in header information and questions section) with answers to the following questions.

• What affect do you expect the choice of X meters to have on the speed of your program? How significant is this affect (if any) on the speed of your algorithm? (E.g., changing it from 15 to 7 meters will make it (a little bit faster)/(a little bit slower)/(twice as fast)/twice as slow).
• Suppose the data file has K unique trips, L unique commutes, and M unique routes.
  • Which will have the largest affect on the speed of your program?
  • Which will have the least affect on the speed of your program?
  • Which (if any) of K, L, and M can be ignored when calculating O() for your algorithm?

Each pair should submit one lab report. Your report should include:

• Complete description of your algorithm. Be sure to discuss other alternatives that you considered and why you chose your method over other alternatives.
• Time complexity analysis for your algorithm.
• Original answers to above questions and any additional discussion.
• Results of your program for the smallgps.zip and biggps.zip files using X = 15 meters and X = 2 meters.
• Benchmarks for your algorithm for the given data files using X = 15, 10, 5, and 2 meters.
• Reactions/suggestions (optional)
• CSP spreadsheet (Continue with the spreadsheet you used for lab 2, but use the naming convention: 385MSOEloginL3.xls.)
• Documented source code
• Results associated with the biggps.zip file are not required; however, please include them if you do have these results.

The first lecture of week 9 will be devoted to student presentations.

• Each team should summarize the results of their solution to lab 3, explain the algorithm they used, and, time permitting, describe the alternatives they considered.
• Presentations are limited to five minutes in duration. Presenters exceeding this limit risk being unceremoniously cut off. A short period for questions and comments from class members will be provided and is not included in the five minute limit. The duration of this period is at the instructor's discretion.
• The presenters should arrange for any necessary audio-visual equipment. The instructor will consult on this process if requested.
• Grading will be based on the likelihood that I would hire your team to work for me on a similar project. (Think of this as a presentation given as part of an interview process.)
• You should take this assignment seriously as it accounts for 40% of your grade for this lab assignment. Students who do not devote adequate time to preparing their presentation should expect to receive a poor grade.

As with any report you submit, correct spelling and grammar are required. In addition, your report should be submitted electronically following the Electronic Submission Guidelines. (You may wish to consult the sample report before submitting your report.) Be sure to keep copies of all your files, in case something gets lost.

If you have any questions, consult the instructor.

© 2001-2006 Dr. Christopher C. Taylor

Office: CC-36C

Phone: 277-7339