Gsoc 2013 Razequl Islam - pgRouting/pgrouting GitHub Wiki

Report List

Biography

  • Name: Khondoker Md. Razequl Islam
  • Country: Bangladesh
  • School and degree: University of Dhaka & Master of Science
  • Major: Computer Science & Engineering
  • Email: [email protected]

GSoC project

My Status Reports

OSGeo Wiki notes: http://wiki.osgeo.org/wiki/Google_Summer_of_Code_2013_Ideas#What_to_expect_during_the_summer

Report 14

Plans

  • Complete integration with pgRouting.

Tasks Done

I have completed the codes for integrating my implementation in pgRouting. I have also checked in the code in the git repository. I am now working with my mentors to build the code with full pgRouting. I am also working on improving the documentation in the meantime. I think this first version of VRP will be ready before the pencil down date. I have enjoyed the summer working with my mentor. I have added time windows in my implementation which was a nice to have feature in the initial plan. Though there are still scopes to improve the solution, I think this is a nice start. I will continue my contribution even after the gsoc is over. I would like to thank my mentors for their excellent support throughout the time. It was a really nice experience.

Next Week Plan

  • Build with pgRouting and update the documentation.
  • Wrap up the first version and submit code samples to Google.

Report 13

Plans

  • Complete integration with pgRouting.

Tasks Done

I have started work on integrating my implementation with pgrouting. I have contacted with my mentors and working as their suggestion. Hopefully I will be able to complete the integration within the soft pencil down. I will fix small issues and work on documentation the rest of the week.

Next Week Plan

  • Complete integration with pgRouting.

Report 12

Plans

  • Add some test cases to test using input from file.
  • Integrate the engine with pgrouting.

Tasks Done

I have added tested the code with some test case. I have added one benchmark test. My implementation is giving the correct answer, but it is not as optimal as the best known result. For example with the added test the best known result can serve all the customer using 15 vehicles where my algorithm needs 18. There are scopes of updating my solution, but I have started to integrate it with pgRouting as this is higher priority. Once it is done, I can review and improve the algorithm.

Next Week Plan

  • Complete integration with pgRouting.

Issues

  • I will communicate with my mentors to find out the best way to integrate it with pgRouting.

Report 11

Plans

  • Test the implementation taking input from files.
  • Document the project in the wiki.

Tasks Done

-I have almost catch up with my plans. The main VRP engine is ready. Soon I will put some test cases in the git for the community to test the engine. The only thing is now I am getting two weeks instead of three for integrating with pgrouting. I hope I can finish it before the soft pencil down date. I am working hard to make it. I still need to do some documentation but I think finish the code is the first priority. But I will do the basic things like read me before I invite the community to test my program.

Next Week Plan

  • Add some test cases to test using input from file.
  • Integrate the engine with pgrouting.

Report 10

Plans

  • Fix small errors.
  • Test the implementation taking input from files.
  • Document the project in the wiki.

Tasks Done

-I am still well behind my plans. I got sick and could not work the first few days in the week. I have made some changes in some functions. I could not finish the documentation as I thought. I still hope that I can catch up with extra effort in the coming weeks.

Next Week Plan

  • Test the implementation taking input from files.
  • Document the project in the wiki.

Report 9

Plans

  • Complete coding on main algorithm.
  • Test the implementation taking input from files.
  • Document the project in the wiki.

Tasks Done

-The VRP Solution is almost ready for input from test files. I am having some runtime errors that I am working on. Once these are fixed, we can check the algorithm with different input from files. I hope this will be finished this week. From the next week I shall start work on integration it with pgRouting.

Next Week Plan

  • Fix small errors.
  • Test the implementation taking input from files.
  • Document the project in the wiki.

Issues

  • I could not work full week in the last week due to Eid. I am back and working in full swing from last Tuesday.

Report 8

Plans

  • Complete coding on main algorithm.
  • Test the implementation taking input from files

Tasks Done

-I have started working on the test code. It is not completed yet. Hopefully it will be fully working on the next week. I am also doing some documentation. I will update the wiki with it

Next Week Plan

  • Complete coding on main algorithm.
  • Test the implementation taking input from files.
  • Document the project in the wiki.

Issues

  • I am with limited Internet access due to Eid. So, communication become less frequent. I will be able to do regular communication from the next Monday.

Report 7

Plans

  • Complete coding on main algorithm.
  • Test the implementation taking input from files

Tasks Done

  • This week we had the midterm evaluation. I have completed the evaluation form. Unfortunately, I am behind my target during the midterm evaluation. I still need some time to make the code work with a test program. Besides this, I am making a document on my project explaining the tasks and approaches I have taken so far. I should have done it earlier. Not doing so is a mistake I have made

Next Week Plan

  • Complete coding on main algorithm.
  • Test the implementation taking input from files.
  • Document the project in the wiki.

Issues

  • Looks like I am bit behind the schedule. I am putting extra effort to get it done in time.

Report 6

Plans

  • Complete coding on main algorithm.
  • Test the implementation taking input from files

Tasks Done

  • I am still coding the main algorithm. Specially I am implementing the Tabu search portion. There are more works than I have initially thought, so it will need some more time to finish to code the main algorithm. I aimed to complete it within mid term evaluation so that it can be tested using input from file. I am working hard to achieve that

Next Week Plan

  • Complete coding on main algorithm.
  • Test the implementation taking input from files

Issues

  • Looks like I am bit behind the schedule. I am putting extra effort to get it done in time.
  • I have uploaded the current code under the branch gsoc-vrp. Though it is not ready to work now, but it will help my mentors to check and evaluate my progress.

Report 5

Plans

  • Continue coding on main algorithm.

Tasks Done

  • I am still coding the main algorithm. This is a Tabu search based algorithm for VRP solution. Right now I am at the middle of the implementation. Hopefully I will be able to complete it this week. Then I will do some test taking input from files. Once it gives desired output, I will work on the wrappers to get the data from database and using other queries.

Next Week Plan

  • Complete coding on main algorithm.
  • Test the implementation taking input from files

Issues

  • There is not much issue right now. I am busy with implementation, so the communication with the mentors has become less frequent. I will need help when I work with distance matrix,

Report 4

Plans

  • Start coding on main algorithm.

Tasks Done

  • Started coding on the main algorithms. The codes for populating the data structures are done.

Next Week Plan

  • Continue coding on main algorithm.

Issues

  • There is not much issue right now.

Report 3

Plans

  • Finalize the architecture and data structure.
  • Start implementing the class skeleton.

Tasks Done

  • Completed primary class designs.
  • Developed the class skeleton. This can be found here

Next Week Plan

  • Start coding on main algorithm.

Issues

  • There is not much issue right now.

Report 2

Plans

  • Finalize the architecture and data structure.
  • Start implementing the class skeleton.

Tasks Done

  • Completed primary class designs, I need to review them with my mentors.

Next Week Plan

  • Implement the class skeletons.
  • Start coding on main algorithm.

Issues

  • I could not progress as I have expected this week due to some unavoidable personal issues. I will put extra effort to catch up this week.

Report 1

Plans

System study, architecture and data structure design.

Tasks Done

  • Started to design the high level architecture and data structure.
  • Discussed various aspects of the project with mentors.

Next Week Plan

  • Finalize the architecture and data structure.
  • Start implementing the class skeleton.

Issues

  • I am having regular communications with the mentors, in most cases I get feedback before I expect to get. So most of the issues are resolved almost instantaneously.