how to model this

Dear list,

I met with a business case and feel frustrated to model it,scenario is like
this

1. logistics department running
2. a lot of customers place orders and goods need to be send to their site
(nationwide)
3. order received and then choose from a list of truck carriers to deliver
goods,a cheapest one should be selected but might not available for the
limit of order-shipment time,so a more expensive one has to be used.

this logistics department run the process on case-by-case basis ,no forecast
from sale department and no bargain power against carriers. I have past 2
years order history and freight history,on which some descriptive statistics
is done and result is really ugly,expensive carriers were used
extensively.I'm trying to find some ways to benchmark or optimize the
process,but due to my lack of statistics modeling skills( I'm from
accounting background and luckily because I have used R to show some
improvement in past analysis,I'm asked to deal with this :(
really can't find the clue where to start.I'll very appreciate if anyone can
point  me to the right direction(books ,papers, or similar situations might
be examined many times) .

thanks

       [[alternative HTML version deleted]]