ago
0 like 0 dislike
0 like 0 dislike
Business Problem: optimize the storage of items between an automated pick area and a manual pick area of a warehouse. The idea is to move items to the automated area right before their demand picks up and keep them stored in the manual area when there is little or no demand.

​

Company Info: Large retail company mainly operating in opportunistic buys (meaning we buy our items in bulk usually out of season, store them until they are in season, then advertise them on the website and sell). This means we hold a lot of inventory that there is no demand for at the moment. Also, we buy new items every year so yearly historical data is not helpful as we only have 15% of repeat buys. This means that we are looking at data sets in a 2-3 month range usually when looking at a unique SKN level.

​

Data set: I have a 6 week time period of all sales on a daily level. The data set includes all relevant order and SKN info. I can get larger time periods as needed.

​

Objective: Create a forecasting tool that can predict the demand for each SKN on a weekly or Bi-weekly level as we will be doing replenishments to the automated pick area each week or every other week. We need to make sure that the units available in the automated pick area are enough to support that weeks demand (or 2 weeks of demand if going with bi-weekly method).

​

Issues: Demand data at SKN level is very sporadic and hard to forecast with. I can use yearly historical data but as mentioned before, the item catalogue differs greatly each year. I can aggregate the demand into item departments and then do a bucket conversion to apply that departments demand to all items that fall into that department. The departments do not change year over year.

​

So with this information, I'm thinking of doing two methods. First being the yearly long term forecasting using the department demand  data then applying to each SKN within the department. Secondly I want to do a short term demand forecast that uses either 2-6 weeks trailing data to predict the next week or 2 weeks demand.

Question: What forecasting methods/tools would be best appropriate for each method? Its been a minute since I did something like this so forgive me. What something like this fall more into a Poisson distro vs a normal? Thank you!!!
ago

No related questions found

33.4k questions

135k answers

0 comments

33.7k users

OhhAskMe is a math solving hub where high school and university students ask and answer loads of math questions, discuss the latest in math, and share their knowledge. It’s 100% free!