inventory, supply chain, procurement management and computer science- General, high-level question [closed]

Question

I would like to ask a rather, general, high-level introductory kind of question regarding inventory management.

So, I was wondering if anyone on SO had any experience/knowledge, or worked with in the past in inventory, supply chain, procurement management settings. What typical problems or challenges one might find in this field and how computer science, mainly algorithms, data structures and optimization can can be employed to deal with such challenges/problems?

Could this be relevant to operational research, queue theory etc? I am not directly related to this field but would need to know how CS is applied in these domains.

An internet search produces some vague results, so I would greatly appreciate any prior-experience insight, educated advice, specific online resources, or even examples. I hope it is ok to ask such a high level question here.

Many thanks in advance

Answer 1

I have some experience with warehouse management systems. Much of it is not very sophisticated in a CS point of view, but there are some juicy optimization problems where CS can be applied. For example, to reduce the time spent to "pick" an order (go through the warehouse and collect the goods for an order), it's desireable to find the shortest way to go to all those places in the warehouse, which boils down to the "traveling salesman problem".

Another place where CS is applied is inventory taking; there are some very clever software products (e.g. INVENT Xpert) that allow a random sample inventory taking to reach the accuracy required by law; this means that instead of going to all storage locations and count the quantity stored there, only a small percentage (5-10%) of the locations are actually counted.

Answer 2

This is a very general question, You probably need knowledge in distributed computing (depends on how big is your operation), Replicating Databases, some knowledge in traveling agent kind of problems and who knows better than you - what else - it is very dependent on the problem you need to solve.

I think you should explain the purpose of the question - so we can narrow the answer to something that might be helpful...

there are also many Of-the-shelf products (that requires a lot of customization, but holds most of what you need in this field).

Answer 3

"What typical problems..."

It is very common to have multiple sites/terminals updating a specific database row/record at the same time so you have to be absolutely bulletproof in your row/record locking and update procedures or you will lose both money and customers. Database concurrency issues are significant and your fail-over systems have to work.

Test under real load. If you expect to have 50,000 different widgets in your warehouse and you expect to have days (day after Thanksgiving) when you get 6,000 hits a second for 9 hours on a particular widget then that's what you test - real data and real volume and at the end of your tests your item quantity, turn, and back-order counts can't be off by even one.

Make sure you've addressed these two issues and you're on your way to a trustworthy system.

Question why are you thinking of writing your own system rather than adapting one that is commercially available?