Strip packing

As of v1.1 the library supports solving strip packing instances (no guillotine cuts).

If you want to solve a strip packing problem, simply instantiate HyperPack with a strip_pack_width integer argument, instead of containers argument.

Beware that providing a containers argument would raise a ContainersError.

Also trying to change the containers attribute in any way in strip packing instances, raises an error.

Now the problem instance is –exclusively– a strip packing problem.

>>> prob = HyperPack(items=items, strip_pack_width=10)
>>> # a strip packing width problem is ready to be solved

Now calling hypersearch or local_search will do the following:

Will try to pack all the items to the smallest possible height.
Every new better solution will push the imaginary container’s height to the height of the new better accepted solution.
Every solution that doesn’t include all the items given will not be accepted.

The above restriction of the container height according to 100% items inclusion can be overridden throught the container_min_height attribute:

>>> prob.container_min_height = 10 # int > 0

Now solving will bypass the total items inclusion restriction and pushes the imaginary container to smaller heights if a better solution is found, with a minimum height restriction of 10.

Reducing the container’s height forces packing to better solutions due to the Point Generations mechanics and the nature of the local search.

Note

The imaginary container’s height is the height of the strip used for packing.

Concluding, if a better utilization of the strip is wanted for trade of number of items included in packing, push strip packing to smaller heights utilizing the container_min_height attribute.

Warning

After solving, the container’s height and container_min_height remains the same from previous solutions, with the exception of multiprocessing hypersearch.

If you want to reset the strip’s height to initial and container_min_height to None, call reset_container_height:

>>> prob = HyperPack(**problem_data)
>>> prob.settings["workers_num"] == 1
True
>>> prob.container_height
200
>>> prob.container_min_height = 100
>>> prob.hypersearch()
>>> prob.container_height < 200
True
>>> prob.reset_container_height()
>>> prob.container_height
200
>>> prob.container_min_height == None
True
>>>
>>> # now with multiprocessing ------
>>>
>>> prob = HyperPack(**problem_data_mp)
>>> prob.settings["workers_num"] > 1
True
>>> prob.container_height
200
>>> prob.hypersearch()
>>> prob.container_height == 200
True

Of course another way would be explicit assignment to container_height and container_min_height attributes.

Apart from the above differentiations, the figures, logging, containers and items structure, remain the same.