Crew Trip Solver Example - Brief Usage

Demo solver

Default settings - Simulated Annealing

Under Quzzi/src you will find test_solver.py. The default configuration is to use Simulated Annealing.

Run as:

  python test_solver.py

Run using Quantum Annealing

Open test_solver.py and go to the end of the file to find the Benchmark loop. Comment out the Simulated Annealing call and uncomment the Quantum Hybrid call as so:

benchmarks = []
for m, model in enumerate(models):

    B = bench(model.A,len(benchmarks))

    # Default uses Simulated Annealing 

    #B.add("Simulated-BQM "+str(m+1), solvers=["neal"], reads=np.linspace(1000,1000,1))

    # To use the Hybrid Quantum Solver, uncomment the line below.
    # Note that the benchmark tool in this loop retains the best solution for any
    # solvers used. Therefore, to ensure getting details of the Hybrid solvers regardless
    # of the outcome, comment or remove other solvers from the benchmark
    
    B.add("Quantum-BQM "+str(m+1), solvers=["hyb"], time=np.linspace(10,10,1))
    
    benchmarks.append(B)
    
    #B.logClear()

Save the file and run as previously shown.

Default use case - 24 Flight segments to sequence

The default use case (datasets/DS2b.csv) contains 24 flight segments per day.

Each row is a flight segment requiring a flight crew.

The problem is to sequence all flight segments into legal trips.

A trip must start at the crew home base (LCA) and finish with a segment that returns home. Transiting through the home base is allowed in this example.

row	fid	Flt	Dep	Arr	dDay	dTime	aTime	aDay	FT	GT
0	001	210	LCA	CWL	d01	0325	0835	d01	ft 0510	gt 0000
1	002	102	LCA	ATH	d01	0405	0550	d01	ft 0145	gt 0000
2	003	602	LCA	BEY	d01	0410	0455	d01	ft 0045	gt 0000
3	004	603	BEY	LCA	d01	0540	0630	d01	ft 0050	gt 0000
4	005	103	ATH	LCA	d01	0635	0810	d01	ft 0135	gt 0000
5	006	122	LCA	CHQ	d01	0725	0905	d01	ft 0140	gt 0000
6	007	302	LCA	CDG	d01	0730	1200	d01	ft 0430	gt 0000
7	008	104	LCA	ATH	d01	0740	0925	d01	ft 0145	gt 0000
8	009	112	LCA	SKG	d01	0910	1110	d01	ft 0200	gt 0000
9	010	211	CWL	LCA	d01	0935	1430	d01	ft 0455	gt 0000
10	011	123	CHQ	LCA	d01	0950	1125	d01	ft 0135	gt 0000
11	012	105	ATH	LCA	d01	1010	1145	d01	ft 0135	gt 0000
12	013	113	SKG	LCA	d01	1155	1350	d01	ft 0155	gt 0000
13	014	206	LCA	BHX	d01	1220	1705	d01	ft 0445	gt 0000
14	015	106	LCA	ATH	d01	1245	1430	d01	ft 0145	gt 0000
15	016	303	CDG	LCA	d01	1300	1700	d01	ft 0400	gt 0000
16	017	202	LCA	STN	d01	1450	1935	d01	ft 0445	gt 0000
17	018	107	ATH	LCA	d01	1515	1650	d01	ft 0135	gt 0000
18	019	108	LCA	ATH	d01	1520	1705	d01	ft 0145	gt 0000
19	020	109	ATH	LCA	d01	1750	1925	d01	ft 0135	gt 0000
20	021	207	BHX	LCA	d01	1805	2220	d01	ft 0415	gt 0000
21	022	604	LCA	BEY	d01	1925	2010	d01	ft 0045	gt 0000
22	023	203	STN	LCA	d01	2035	0050	d02	ft 0415	gt 0000
23	024	605	BEY	LCA	d01	2055	2145	d01	ft 0050	gt 0000

Output - Trip sequences

Segments will be sequenced into “trips”, each originating and terminating at the crew base (LCA in this use case).

The output will look like this:

Allocated 3746 qubo variables 0 to 575
          24 start flags 576 to 599
          3146 additional variables 600 to 3745
1 models
1 benchmarks
Solving Profile: Quantum-BQM 1    Solver: hyb      Time: 10.0 -- Reads: 100 -- Chain: 1
Solving using the LeapHybridSolver...(time limit is 10.0 )
Energy : -480006.99227539264
Trip
0 007 302  LCA CDG d01 0730 1200 d01 ft 0430 gt 0000 0.0
1 016 303  CDG LCA d01 1300 1700 d01 ft 0400 gt 0100 0.0
Trip
2 002 102  LCA ATH d01 0405 0550 d01 ft 0145 gt (1255) 0.0
3 005 103  ATH LCA d01 0635 0810 d01 ft 0135 gt 0045 0.0
4 009 112  LCA SKG d01 0910 1110 d01 ft 0200 gt 0100 0.0
5 013 113  SKG LCA d01 1155 1350 d01 ft 0155 gt 0045 0.0
6 017 202  LCA STN d01 1450 1935 d01 ft 0445 gt 0100 0.0
7 023 203  STN LCA d01 2035 0050 d02 ft 0415 gt 0100 0.0
Trip
8 001 210  LCA CWL d01 0325 0835 d01 ft 0510 gt (2125) 0.0
9 010 211  CWL LCA d01 0935 1430 d01 ft 0455 gt 0100 0.0
10 019 108  LCA ATH d01 1520 1705 d01 ft 0145 gt 0050 0.0
11 020 109  ATH LCA d01 1750 1925 d01 ft 0135 gt 0045 0.0
Trip
12 003 602  LCA BEY d01 0410 0455 d01 ft 0045 gt (1515) 0.0
13 004 603  BEY LCA d01 0540 0630 d01 ft 0050 gt 0045 0.0
Trip
14 015 106  LCA ATH d01 1245 1430 d01 ft 0145 gt 0615 0.0
15 018 107  ATH LCA d01 1515 1650 d01 ft 0135 gt 0045 0.0
Trip
16 006 122  LCA CHQ d01 0725 0905 d01 ft 0140 gt (0925) 0.0
17 011 123  CHQ LCA d01 0950 1125 d01 ft 0135 gt 0045 0.0
18 014 206  LCA BHX d01 1220 1705 d01 ft 0445 gt 0055 0.0
19 021 207  BHX LCA d01 1805 2220 d01 ft 0415 gt 0100 0.0
Trip
20 008 104  LCA ATH d01 0740 0925 d01 ft 0145 gt (1440) 0.0
21 012 105  ATH LCA d01 1010 1145 d01 ft 0135 gt 0045 0.0
Trip
22 022 604  LCA BEY d01 1925 2010 d01 ft 0045 gt 0740 0.0
23 024 605  BEY LCA d01 2055 2145 d01 ft 0050 gt 0045 0.0
----------------------------------

Objective - Minimize non-flight time spent by crew members

The objective (above and beyond basic constraint compliance for the structure of trips) is to minimize the crew duty time spent that is not directly for performing a flight. In other words

Minimize Sum (Duty Time minus Flight Time) for all trips.

More precisely: Report time of 60 minutes and Release time of 30 minutes for each trip Connection time (ground time) between flight segments within each trip

The number of resulting trips is not set ahead of time.

Note that flight time limitations per trip are not applied in this model.