0

Have to solve a distribution optimization problem in R but can't seem to bring my head around the approach for formulating it effably:

A is a matrix of accounts and account holders with the relevant amounts available to each coordinate. The entire matrix is a network with n:m relations between accounts and account holders. I need to optimize the distribution of amounts under the following two constraints:

  • C1 => all account holders have to be covered by at least 100k (or at least the max amount possible) / this condition has priority
  • C2 => maximize the amount on accounts that can be fully covered (i.e. it's better to cover only one account to 100% instead of multiple accounts if the total sum is higher on the single account)

We don't have any negative values involved.

I have tried to approach the problem using OptimSimplex and GA but stumbled about getting the two constraints to work on the same distribution yet including a priority for the first.

manlio
  • 18,345
  • 14
  • 76
  • 126
Tom Debus
  • 23
  • 1
  • 5
  • C2 sounds like an objective function, not a constraint. What is your objective function? – C8H10N4O2 Jun 26 '15 at 13:38
  • also, what's with the genetic-algorithm tag, this looks like an integer linear program? – C8H10N4O2 Jun 26 '15 at 13:39
  • Hi Manilo, thought to solve it using GA - an you're right - C2 is really the objective function but it can only be achieved once the constraint of C1 is fulfilled. But how you fulfill C1 impacts in turn how & what you can optimize. – Tom Debus Jun 27 '15 at 18:49

0 Answers0