Computational complexity is a problem faced in the frequency spectrum auction and is called the Winner Determination Problem. The complexity increases when the object auction increases and partitioned into groups of objects based on some pre-defined zone. With the zoning scheme, there are three alternative determination of winners, the bidders get the whole object of auction at the entire zone (also called all or nothing scheme), auction bidders get the whole object in some specific zones (also called partially zone scheme), bidders are assigned some object in some zones certain (partially block zone). In this paper the issue is viewed as a matter of Knapsack auctions, both Knapsack 0/1 and Bounded Knapsack. The issue is then simulated for the three alternatives to obtain the best solution based on the consideration of the highest revenue. We uses 14 data sets for simulation, with the technique used is the Brute Force and Dynamic Programming which has been implemented in the Java programming language. For the same data sets, the simulation results establish that the partially zone is producing the highest revenue, followed by partually block zone scheme, then the all or nothing scheme