You're right about that.and don't get me wrong. Feasibility shouldn't be an issue.any random selection that gives each location its max, regardless of value, is feasible.not sure how to proceed at this point. When I run solver, it spins for a while (using the evolutionary engine) or a very short while (using the GRG non-linear) and then tells me it cannot find a feasible solution. Those maximums add up to 182, so every vehicle should be allocated to a location. The constraints are then that a) the decision variables must be integers and b) each of the five locations has a maximum number of vehicles it can support. The column labeled Average is the average value for all 182 decisions.this is the objective cell to maximize. The column labeled value is then a sumproduct of the value columns and the 0/1 location columns.basically just gives you the value for the chosen location. The next five columns are just IF statements so that there will be a 1 in the column associated with the chosen location and 0's elsewhere. The column labeled Choice is the decision variable (changing cell) in the solver model. Column 1 is a vehicle ID, the next five columns are the "value" associated with placing that vehicle at location 1, location 2, etc. Each row represents a vehicle that may be placed at one of five locations. Although it goes on.in total there are 182 rows of data.
0 Comments
Leave a Reply. |