Travelling Thief Problem (TTP) is a hybrid of the classic NP-Hard Travelling Salesman Problem (TSP) and the Knapsack Problem (KP). DynTTP is an extension of the classic TTP where the problem constraints can change during runtime.
Conference: Accepted at ICONIP'20
Paper: pdf
Authors: Ragav Sachdeva, Frank Neumann, Markus Wagner
Compile
g++ -std=c++11 dttp.cpp -o dttp
Run
./dttp instance solver num-runs disrupted-entity percentange-toggle reoptimiser num-disruptions frequency tourfilepath(optional)
e.g.
./dttp instances/a280_instances/a280_n279_bounded-strongly-corr_01.ttp S2 30 items 0.3 viaBitflip 10 1 tours/a280/
Running the above will produce a .txt file in the output folder containing all the objective values and the corresponding eval count in an array-like format.
We are attaching a jupyter notebook for visualisation that shows how to read data from the output .txt files.
You can find all the plots for our experiments under the Results folder.