Crazy question re optimization
Posted by Mike Fleuette on August 30, 2001 12:36 PM
I am trying to staff a number of short-term projects and am looking for a way to optimize the way people are selected for different roles. The goal is to maximize the score for the team by slotting each person into one role such that the sum of (aptitude*priority) is maximized. There will be more staff than roles (Hopefully!). An example of the concept is below; the obvious question is how do you identify the best person for a given role to maximize the effectiveness of the team?
intersection (Role/Person) = aptitude
Priority 1 3 2
Role --> x y z
Person a 0.4 0.5 0.8
| b 0.2 1 0
| c 0.3 0.4 0.3
V d 0.8 0.5 0.5
y ?
z ?
x ?
Score ? (maximized)
Posted by Mike Fleuette on August 30, 2001 12:37 PM
Posted by Eric on August 31, 2001 8:31 AM
WOW, that's an incredibly complex problem!
I'd check the "traveling salesman problem" or TSP for short. There's a good jumping-off point at
http://www.math.princeton.edu/tsp/
What you're asking for is, I think, a bit less simple than it first appears, and there is no readily available excel tool that can handle it. Although I'm sure you could find an excel-buff who could implement some of the models used to solve these problems (e.g., "random walk" solution from ant-based analogies).
Good luck!
