DA-Project2

Compile the performance test (in code directory)

g++ -std=c++17 -I. performance/PerformanceRunner.cpp FileReader.cpp Truck.cpp Algorithms.cpp Pallet.cpp -o performanceRunner

It runs the the follow datasets: datasets = {"01", "02", "03", "07", "08", "10", "11", "12", "13", "14"};

Run the ilp knapsack python script

python3 code/ilp_knapsack.py input/dataset/TruckAndPallets_01.csv input/dataset/Pallets_01.csv output/output.txt

Run main code (in code directory)

g++ -std=c++11 main.cpp -o main Algorithms.cpp FileReader.cpp Menu.cpp Pallet.cpp Truck.cpp PerformanceEvaluator.cpp

Run the graphic script (in project directory)

• has to have a virtual environment (venv) with pandas and matplotlib installed

source venv/bin/activate
python3 code/performance/plot_runtime.py

Input files - use like this

• TruckAndPallets_01.csv
• Pallets_01.csv

Implemented datasets

Pallets_11.csv and TruckAndPallets_11.csv

• Tie-Breaker and Greedy Trap
• Multiple subsets yield the same total profit (e.g., [1,2], [3,4,5,6])

Pallets_12.csv and TruckAndPallets_13.csv

• Stress DP without Brute-Force Explosion
• n = 12 (safe for brute-force)
• capacity very large - W = 50000

Pallets_13.csv and TruckAndPallets_14.csv

• Stress Greedy Algorithm
• Using W = B, v1 = 1, w1 = 1, v2 = B - 1, w2 = B

Pallets_14.csv and TruckAndPallets_14.csv

• Backtracking Efficiency test
• Many combinations exceed the weight limit quickly, allowing backtracking to prune branches early