#Beginner, looking for advice regarding the appropriate type of AI model to solve a specific problem

1 messages · Page 1 of 1 (latest)

small wedge
#

Hi!
I just want to preface this by saying that I'm not looking for anybody to solve my problem. I've just never messed around with AI before and I don't know up from down - if someone could point me in the general direction of the type of AI model that is optimal for this sort of work, I can learn everything myself from there.

I have 5 days to make an inventory management and automatic reordering system, which takes in a bunch of data points such as:

  • current stock amount
  • rate of usage
  • time to receipt of a new order of supplies
  • reliability of specific suppliers (i.e. how long they take to actually deliver supplies versus their delivery estimation)
  • historical data about the usage rate of specific items around certain times of year
    and make a system that automatically decides how much of a stock to order, and when to order it, to ensure optimal storage costs and no shortages.
indigo mason
#

https://en.m.wikipedia.org/wiki/Linear_programming

Linear programming

Linear programming (LP), also called linear optimization, is a method to achieve the best outcome (such as maximum profit or lowest cost) in a mathematical model whose requirements and objective are represented by linear relationships. Linear programming is a special case of mathematical programming (also known as mathematical optimization).
Mor...

small wedge
#

Alright, I'll look into this

#

Thank you:)

indigo mason
#

No worries

#

I should justify my response

#

This problem can be solved with perfect accuracy using linear programming. However, traditional methods like this are intractable / impossible to get right for complex pattern recognition problems: vision, language, sound, driving, etc.

Machine learning is hard to design and configure. However, it really excels in complicated problem domains, because it learns to extract patterns from unstructured data itself.

That suggests a split: try to use traditional methods when the problem and data are highly structured. Reach for ML when things get more confused, unclear, or complex.

#

You can identify this split by considering the sort of problems big AI companies focus on.

small wedge
#

okay

#

I'll keep that in mind for the future

#

that's super helpful

small wedge