GSoC 2014 Mukul Priya R7

What did i do this week ?

• As mentioned in the new proposed time line , i worked on reducing the number of vehicles of the generated initial solution before feeding this solution into the optimizer.However This procedure should work better after the optimizer has done its work.

What will I be working on next week?

• I will be working on the optimizer this week as mentioned in the plan. I will be implementing the 2-opt* method for generating neighbourhood inside Tabu Search.

Did I meet with any stumbling blocks?

• The method adopted by me to reduce the vehicles was easy to implement , however the results were not what i expected , i think after we have an optimal solution It should work .However I will give some more time to it during this weekend and see if i can improve it.

Goals and New Timeline

• week 7: Try and reduce the number of vehicles form the initial solution and then capitalize on the outcome and feed it into the optimizer .

• week 8: Implement the 2-opt* (exchange/removal-insertion of orders between two routes ) , observe the initial results and see how it can be further improved . I have already implemented most of the portion of this method .

• week 9-10: More improvement(after modification )if possible and Integrate it with the code and test it for CVRPTW (build test suites if required).

• week 11-12: Tweak the optimizer for CVRP (without time windows)

• week 13-14: Documentation and integration with pgRouting .

• Regarding community discussions , as discussed with my mentors all the technical discussion involving the project will now be done on the developers mailing list.This will be helpful in eventually informing the community about the problems that i may face and in case of that they will also be able to give some suggestions.