kiwicom / travelling-salesman

Rules for Kiwi.com travelling salesman competition

Rules:

• Find the Hamiltonian cycle in time dependent flight graph
• You are given:
  • m flights flying between n cities
  • n days to visit n cities
  • Start/end airport (round trip)
• Being at day i in city x means you can catch all the flights on day i+1 from city x
• All flights are immediate, they take no time.
• On each day you have to board exactly 1 flight.
• On some days, there might not be a flight between some cities.
• We guarantee, that at least one Hamiltonian cycle in any given graph, exists.

Evaluation

• You have 30 seconds for each dataset to produce valid output. After this period we kill your program if it hadn't already finished and process what your program outputted up to this point. Invalid output will result in 0 points but killing your program as such doesn't affect awarded points as long as you manage to produce valid result. Detailed description of our scoring system can be found in scoring README.
• Evaluation data will be sampled the same way as the training real data - we will use real world prices of flights.

Program:

• Executable in /app/run
• Input will be STDIN, output will be STDOUT.

Input, STDIN

<City to start from>  # always a single IATA code
<FROM> <TO> <DateOfDeparture> <PRICE>
<FROM> <TO> <DateOfDeparture> <PRICE>
...

Some remarks about input:

• <FROM>, <TO> - IATA codes - 3 English alphabet uppercase letters A-Z.
• <PRICE> will be positive integer that fits into uint16_t - 16 bit unsigned integer (0 < <PRICE> <= 65535). (No prefixes, no hexa, nothing. Just number. :-])
• Input data will be ordered lexicographically (FROM, TO, DateOfDeparture), with FROM/TO ordering to be also lexicographical. If this seems to be confusing, just look at the output of data_generator.py.

e.g.:

NAP
NAP BRQ 0 10
FCO BRQ 0 641
BRQ FCO 1 40
BRQ NAP 1 510
FCO NAP 2 3

Output, STDOUT

This will be strictly validated for syntax, and against input.

Format

• We expect n+1 lines.
• First line will be total price.
• Second to last(n+1) line will contain flights in the same format as the input.

e.g.:

53
NAP BRQ 0 10
BRQ FCO 1 40
FCO NAP 2 3

travelling-salesman's Issues

Add uncompressed versions of data

The .zip versions of data are harder to use. As GIT does compression by itself, I see no reason why not replace them with uncompressed .txt?