WebCab Optimization for COM v2.6

LinearProgramming.DualSensitivity Method 

Performs sensitivity analysis of the primal problem based upon the dual variables.

public double DualSensitivity(
   double[] dualSolution,
   double[] boundaryShifts
);

Parameters

dualSolution
The location of the solution of the dual problem which can be found through the application of the method DualSolve. The solution of the dual problem should be provided in the following form: the k-th term of the array should refer to the coordinate value (in the dual coordinates) of the k-th coordinate.
boundaryShifts
An array where the i-th term corresponds the absolute shift of the constant term of the k-th constraint where the constraints are listed with the inequality constraints first followed by the equality constraints.

Remarks

This procedure offers a very efficient means by which to analyze the sensitivity of the extremum of the object function to shifts in the boundaries. In particular, for (small) parallel shifts of the linear boundaries we evaluate the absolute change in the value of the extremum (i.e. maximum or minimum) of the primal problems object function.

Interpretation in term of the Factory Problem Example

Within our factory example the object function represents the profit and the (inequality and equality) constraints which represent the limitations on the means of production such as man hours, raw materials and so on. The sensitivity analysis performed here considers the effect on the maximum profit of the factory when the amount of the means of product represent by the constraints is varied. That is, the constant terms of the inequalities and equality constraints is varied. In the notation given below this corresponds the shifts on the vector b, containing the constraints constants terms.

See Also

LinearProgramming Class | WebCab.COM.Math.Optimization.LinearProgramming Namespace