Shipping Matrix Calculations

[Chris Buckett]

Hi,

I was just wondering if anyone has had any experience in developing a
shipping matrix.

We are based in the UK, and we need to ship both locally and
internationally. Local shipping is priced by weight and doesn't matter
about the number of packages, international shipping is by weight and
package size, so we need to know the number of packages.

Whilst we can decide how many of product x fits in a box, and what each
product weighs, what happens if someone orders:
-2 of product x,
-3 of product y and
-4 of product z
the problem here is how can we tell if it will all fit in one box
etc...?
Should we need to calculate the volumes of each item?

Maybe we should treat each order of each product type as a separate
shipping item, e.g. the example above shipped internationally should be
treated as three shipping orders and priced accordingly (even though
they may all fit in one box).

Has anyone solved this problem in the past?

Thanks for any advice!
Chris.

 

[Bob Powell [MVP]]

Hi Chris,
This forum is more for the use of C# language nuts... You will probably be
better off in one of the SQL or database forums.

--
Bob Powell [MVP]
Visual C#, System.Drawing

Ramuseco Limited .NET consulting
http://www.ramuseco.com

Find great Windows Forms articles in Windows Forms Tips and Tricks
http://www.bobpowell.net/tipstricks.htm

Answer those GDI+ questions with the GDI+ FAQ
http://www.bobpowell.net/faqmain.htm

All new articles provide code in C# and VB.NET.
Subscribe to the RSS feeds provided and never miss a new article.

 

[chrisb]

Hi Chris,
This forum is more for the use of C# language nuts... You will probably be
better off in one of the SQL or database forums.

Ok, I know now that there are probably better groups to post this to.
I just happen to work in C#.
Chris.

 

[Shawn B.]

Should we need to calculate the volumes of each item?

Maybe we should treat each order of each product type as a separate
shipping item, e.g. the example above shipped internationally should be
treated as three shipping orders and priced accordingly (even though
they may all fit in one box).

Has anyone solved this problem in the past?

This makes for a really interesting algorithm; one that I've spent some time
on in the past. If you calculate based on volume alone then you might be
able to decide if you can fit in another area of a larger volume but you
won't be able to determine whether you can snug in the box. For example, a
shipping container might have a volume of 27 sq. meters but the dimensions
might be something like 3 x 3 x 3 meters. The shipping items you are
shipping might be 10 x 1 x 1 meters and won't fix in the shipping container.

Ideally, you'll be using the dimensions of the shipping items and then fit
them together like in the game of tetris. There's many ways to do this and
the way I chose was to create a 3 dimensional matrix and fill the
hypothetical unit with a value to indicate that a shipping item takes up
that space. The size of the matrix will depend on your smallest
calculatable unit (3.8 meters is really 38 centimeters, etc.) so one of your
matrix indices represents 1 centimeter. On another dimension of the
shipping item, perhaps the smallest unit is 1 meter so your other matrix
index represents 1 meter. Of course, the matrix can be huge but like I
said, there might be better ways to do this.

In the end, you can't use volume or dimensions alone to run an abstract
calculation because the math will always work out. You have to somehow be
able to fit them together to see if they will fit in the shipping container
because even though the volume works out and the dimensions seem to, once
you "stack" them together you might not have enough space to fit the
remaining items. Thus, think of it in terms of 3D tetris and you'll do
fine, conceptionally. Another thing is to weed out the largest of the
shipping items, to see if the large of its dimensions exceeds the larger of
the shipping container and then the to the second largest then to the 3rd
largest. If any of them are too large then you automatically know it won't
fit. The other thing is, we keep track of which dimension represents the
"top" of each the shipping item and the shipping container (if there must be
a top such as a television set or something) and that helps to narrow down
the game of tetris, so-to-speak.


Thanks,
Shawn

 

[Shawn B.]

In addition to all I said in the previous post I forgot the mention an
"optimzation" I eventually gravitated towards, but the optimzation will only
work in certain conditions: if you ship the same products most of the time
then you can pre-compute shipping arrangements and re-use them.

For example, you know that 1x Product A will fit nicely into Package A-1.
For all orders of 1x Product A, you use Package A-1.

You know that 1x Product A + 3x Product B + 4x Product C will fit nicely
into Package ABC-134. Whatever Package ABC-134 is, you define it. I can be
a 1 x 3 x 3 meter box. In any case, you've already figured out how to
"snap" the shipping items into the Package and you just record it. On all
future shipments of the same order (or any other order that match the
dimensions of Product A/B/C) you use the same Package ABC-134.

You can either manually enter this data into the system for future use or
you just record it somehow from a previous computation that automatically
computes whether the items will fit in the package and in what optimal
stacking scheme. Thereafter, you just do a lookup and there is no
computation necessary.

There are more factors to consider, like whether Product B is fragile
(glass) and Product C is an ironball, you obviously won't ship them
together. Again, you can create attributes for each package and have some
business rules that dictates what items go together and what items cannot
and then you basically store your information accordingly. Sometimes glass
and iron can ship together. Maybe you have a flat iron plage of some kind
you can just pack the glass on top and indicate to the shipper "this side
up".


Hope I got you thinking.


Thanks,
Shawn

 

[chrisb]

Thanks for that Shawn,
It's good to get another head involved - I was kind of thinking of a 3D
tetris scenario, but I hadn't considered fragile and non-fragile items,
and the concept of products having a "side up."

We had thought about a sort of "self learning" system for calculating
shipping - that is, when someone orders a product mix, we enter the
actual shipping cost to their destination (after having guestimated
shipping first), and next time someone orders the same product mix to a
similar destination, we use the orginal actual shipping cost as a
lookup. I'm just concerned that we may take more time "teaching" the
system about product mixes, and not get enough hits (I guess I'm going
to have to analyse previous orders). I'll look into doing a lookup for
product packing as you suggest and have a chat with the warehouse guys.

At present, we're using a simple percentage of product cost with a
minimum fee, which works out for quite a lot of orders, but we're
starting to sell more high value (but small) items, which makes for
cheap shipping costs for us, but expensive shipping costs for the
customer. Working out shipping for multiples of individual products
seems relatively simple - it's combinations of products that adds the
hard work!

I expect that unless I come up with a better solution, we'll probably
go with a lookup matrix for individual product orders and common
product mixes, and a sort of guestimate (erring in our favour) for
unknown product mixes - possibly with the "learning" facility to
convert an unknown to a known mix.

I've had a quick look at the UPS and FEDEX api's, but they all need to
know the size & weight of the package before giving a price (I guess
that should've been obvious).

Cheers,

Chris.
http://blog.cecatech.co.uk/chrisb

 

[Shawn B.]

If you don't want to spend the effort to dynamically compute packing
strategies (which can be time-consuming and expensive) for what you view as
a marginal benefit, you're better off to go through the pain of figuring out
a strategy when you need to and then "caching" that stretegy for future
lookups as I stated earlier and as you implied you might do. Even for a
place like Amazon, having such a dynamically computing program may not be so
much of a benefit as what they sell are usually of predictable
characteristics and will likely fit into any of various package sizes they
have (I know, I purchased over 500 books for amazon (computer programming
books) of various quantities) and usually its the same box for a certain
quanity.

Some place like Wal-Mart or a freight company, on the other hand, might get
much more utility from such a program and they ship larger objects into more
shipping space and usually a wide-degree of contents that can't easily be
predicted.

It really depends on what you perceive your overall gains to be.


Thanks and Good Luck,
Shawn