My model maximizes the profit over a given number of time periods and one part of the object function sums the revenue of selling assets acquired in an earlier period. When I test the model I find that the probability for the model completing a global search and finding the optimal solution increases when the input sales prize are lowered. For relatively "high" sales prizes, which are still lower than the purchase prize, the model only completes LP relaxation with status: Unbounded. If the sales prize in the last period is set to 0 in the input file then Xpress uses simplex dual algorithm. If it is given a value it uses simplex primal. Since its a maximization problem I don't understand why it becomes unbounded when the selling prize increases and gives a solution when it approaches zero, especially because I have a constraint forcing to sell assets before their lifetime(given in time periods) expires. If anyone would have a suggestion to where the problem might be I would be very grateful.

Without looking at the problem, I would like to suggest a test. Can you change the constraint forcing to sell assets before their lifetime to a preference (where you pay a penalty for not selling assets.). See if the problem is still unbounded. Good luck.

Tauhid