Benchmark Instance Generation¶

MDVRPTW environment instances are introduced in [Vid2013]:

Benchmark instance generation settings are defined in file benchmark_instances_generator.py.

Data files format¶

************************
* MDVRPTW instances    *
************************

The first line contains the following information:

    type m n t d

where

    type = 8 (MDPVRP)

    m = number of vehicles

    n = number of customers

    t = number of days

    d = number of depots

The next t lines contain, for each day/depot, the following information:

        D Q

where

    D = maximum duration of a route

    Q = maximum load of a vehicle

The next lines contain, for each depot and for each customer, the following information:

    i x y d q f a list e l

where

    i = customer/depot number

    x = x coordinate

    y = y coordinate

    d = service duration

    q = demand

    f = frequency of visit

    a = number of possible visit combinations

    list = list of all possible visit combinations

        e = beginning of time window (earliest time for start of service),
            if any

        l = end of time window (latest time for start of service), if any

            Each visit combination is coded with the decimal equivalent of
            the corresponding binary bit string. For example, in a 5-day
            period, the code 10 which is equivalent to the bit string 01010
            means that a customer is visited on days 2 and 4. (Days are
            numbered from left to right.)

BenchmarkInstanceGenerator¶

class maenvs4vrp.environments.mdvrptw.benchmark_instances_generator.BenchmarkInstanceGenerator(instance_type: str = 'Vidal', set_of_instances: set = None, device: str | None = 'cpu', batch_size: Size | None = None, seed: int = None)[source]¶

MDVRPTW benchmark instance generation class.

__init__(instance_type: str = 'Vidal', set_of_instances: set = None, device: str | None = 'cpu', batch_size: Size | None = None, seed: int = None) → None[source]¶

Constructor. Create an instance space of one or several sets of data.

Parameters:

instance_type (str) – Instance type. Can be “Solomon” or “Homberger”. Defaults to “Solomon”.
set_of_instances (set) – Set of instances file names. Defaults to None.
device (str, optional) – Type of processing. It can be “cpu” or “gpu”. Defaults to “cpu”.
batch_size (torch.Size, optional) – Batch size. If not specified, defaults to 1.
seed (int) – Random number generator seed. Defaults to None.

Returns:

None.

get_instance(instance_name: str, num_agents: int = None) → Dict[source]¶

Get an instance with custom number of agents.

Parameters:

instance_name (str) – Instance file name.
num_agents (int) – Number of agents. Defaults to None.

Returns:

Instance data.

Return type:

Dict

classmethod get_list_of_benchmark_instances()[source]¶

Get list of possible instances from benchmark files.

Parameters:: n/a.
Returns:: None.

load_set_of_instances(set_of_instances: set = None)[source]¶

Load every instance on set_of_instances set.

Parameters:: set_of_instances (set) – Set of instances file names. Defaults to None.
Returns:: None.

parse_instance_data(instance_data: list, instance_name: str) → Dict[source]¶

Parse instance data list into a dictionary.

Parameters:: instance_data (list) – Instance data.
Returns:: Parsed instance data.
Return type:: Dict

random_sample_instance(instance_name: str = None, num_agents: int = None, num_nodes: int = None, num_depots: int = None, seed: int = None, device: str | None = 'cpu') → Dict[source]¶

Sample one instance from instance space, randomly adjusting the nodes.

Parameters:

instance_name (str) – Instance file name. Defaults to None.
num_agents (int) – Total number of agents. Defaults to None.
num_nodes (int) – Total number of nodes. Defaults to None.
seed (int) – Random number generator seed. Defaults to None.

Returns:

Instance data.

Return type:

Dict

read_instance_data(instance_name: str) → Dict[source]¶

Read instance data from file.

Parameters:: instance_name (str) – Instance file name.
Returns:: Instance data.
Return type:: Dict

sample_first_n_services(instance_name: str = None, num_agents: int = None, num_nodes: int = None, device: str | None = 'cpu') → Dict[source]¶

Sample first n nodes.

Parameters:

instance_name (str) – Instance file name. Defaults to None.
num_agents (int) – Total number of agents. Defaults to None.
num_nodes (int) – Total number of (n) nodes intended. Defaults to None.

Returns:

New instance of the first n nodes.

Return type:

Dict

sample_instance(num_agents: int = None, num_nodes: int = None, capacity: int = None, service_times: float = None, instance_name: str = None, sample_type: str = 'random', batch_size: Size | None = None, n_augment: int | None = None, seed: int = None, device: str | None = 'cpu') → Dict[source]¶

Sample one instance from instance space.

Parameters:

num_agents (int) – Total number of agents. Defaults to None.
num_nodes (int) – Total number of nodes. Defaults to None.
capacity (int) – Capacity of the agents. Defaults to None.
service_times (float) – Service time in the nodes. Defaults to None.
instance_name (str) – Instance name. Defaults to None.
sample_type (str) – Sample type. It can be “random” or something else for “first n”. Defaults to “random”.
batch_size (torch.Size or None) – Batch size. Defaults to None.
n_augment (int, optional) – Data augmentation. Defaults to None.
seed (int) – Random number generator seed. Defaults to None.

Returns:

Instance data.

Return type:

Dict

sample_name_from_set(seed: int = None) → str[source]¶

Sample one instance from instance set.

Parameters:: seed (int) – Random number generator seed. Defaults to None.
Returns:: Instance sample name.
Return type:: str

References

[Vid2013]

M, T. Vidal, T. G. Crainic, M. Gendreau, and C. Prins, A hybrid genetic algorithm with adaptive diversity management for a large class of vehicle routing problems with time-windows, Comput. Oper. Res., vol. 40, no. 1, pp. 475-489, 2013.