M B A 8590 Lecture 11: Clemson - Decision Modeling - Class Notes - 11 - 4.317 - Jose Pena
Document Summary
Decision modeling 859 class notes - #11. The linear assumption is not used anymore: the tricky part is your solving algorithm isn"t going to give an exact solution but an aprox solution, you wont know if the solution is accurate or no. When one variable is non linear then the model becomes non linear: objective function, or other constraint. In non-linear the model stops when it thinks that is close to the accurate solution but it isn"t always stopping close to the optimal solution. Grg algorithm: is able to solve non lineal programming: recommended to be used before going to a more advanced programming, don"t use it for linear programming, simplex is the option for linear. Local and global optimum in linear they match but not for non-linear. In non linear for example a cubic expression the algorithm may pick one peak rather than the other based on local optimization rather than global optimum.