quantum-stochastic-optimization-railways

Application of quantum computation for stochastic optimization on the example of railway/tramway network in Baltimore.

Files:

1. QTrains - source code

2. tests - testing source code

3. solutions - stored solutions of railway problems, if for particular parameters setting computations have already been stored, new computation will not be performed and the particular file will not be overwritten

4. QUBOs - qubos of railway problems

5. QAOA Results - results of quantum gate computing via QAOA

6. histograms - histograms from data analysis

7. histograms_soft - histograms from data analysis with a soft check of minimal passing time constrain.

Quantum annealing

In process_q_annealing.py trains scheduling problems are solved via Integer Linear Programming and quantum (or simulated) annealing

Arguments:

• --mode MODE: process mode: 0: prepare only QUBO, 1: make, computation (ILP and annealing), 2: analyze outputs, 3: count q-bits, 4: prepare Ising model 5: CPLEX benchmark - by default: 2
• --simulation SIMULATION: if True solve/analyze output of simulated annealing (via DWave software), if False real annealing - by default: False
• --softern_pass SOFTERN_PASS: if true analyze output without feasibility check on a minimal passing time constrain - by default: False

Example usage:

python3 process_q_annealing.py --mode 1 --sim True

Solve the series of problems by simulated annealing (does not perform calculations already performed and saved).

python3 process_q_annealing.py --mode 1

Solve the series of problems by real D-Wave annealing (does not perform calculations already performed and saved).

python3 process_q_annealing.py --mode 2 --softern_pass True

Quantum gate computing

Script process_q_gates.py saves QUBO and the ground state as well as analyses output dedicated to gates computing.

Arguments:

• --notrains NOTRAINS number of trains, 1,2,4 are supported, by default: 2
• --savequbo SAVEQUBO if True prepare qubo else to analyze outputs, by default: False
• --nolayers NOLAYERS number of layers of QAOA in analyzed data, by default: 1
• --datafile DATAFILE file with data, by default: "QAOA Results/IonQ Simulations/"

Example usage:

python3 process_q_gates.py --notrains 2 --nolayer 1 --datafile "QAOA Results/IonQ Simulations/"

Analyzes 2 trains results in "QAOA Results/IonQ Simulations/" where 2 layers of QAOA was used

python3 process_q_gates.py --notrains 2 --savequbo true

Prepared QUBOs for 2 trains problems and save them in QUBOs/gates/2trains/

Preparing plots for the article

Script plots4article.py creates .csv files for high-quality plots for article, and saves them in the article_plots folder.

Funding

Scientific work co-financed from the state budget under the program of the Minister of Education and Science, Poland (pl. Polska) under the name "Science for Society II" project number NdS-II/SP/0336/2024/01 funding amount 1000000 PLN total value of the project 1000000 PLN