We need to structure bounds as array of tuples. After that we can extract our unknown variables and bounds. We will access to model table Variables using “Input” feature and to model table Constraints using “Lookup” feature. Next step is to import all data from our two input model tables.
Transform boundary definition code#
Below you can see code that we implemented in script.įirstly, we need to import all necessary packages: import pandas as pd We use Python RFA to perform linear programming calculations. For constraint type we can choose following values “=”, “>”, “=” and also we can set our optimization problem to minimization (MIN) or maximization (MAX). Name of fields that we use to store coefficients should match name of unknown variables defined in model table Variables. In model table Constraints, we define all constraints as well as objective function: In model table Variables, we define all unknown variables and their lower and upper boundary: In order to explain as simply as possible the way of solving linear programming tasks and the way a Python script works, we will use the following optimization problem: Approach with horizontal structureįor this approach we use 2 input model tables and one Python RFA: Model table with equation in textual format – this approach has same advantages as previous one, but this approach requires very careful definition of equations in text format.Model table with vertical structures – instead of defining input for Python RFA in matrix structure, matrix structure is transformed into vertical structure in order to keep number of columns fixed, but also keep possible to define very complex linear programming problems without restrictions on the number of unknown variables and constraints.Model table with fixed number of columns (horizontal structure) – in situation when we have relatively small number of optimizing variables and this number is fixed.I implemented 3 different input structures for defining LP problems:
You can find more information about linprog function on the following link: For linear programming problems we can use function linprog from scipy.optimize subpackage. Python has a very powerful library called Scipy that allows users to perform very complex calculations and solve optimization problems. In this blog I will describe three approaches for defining and solving Linear Programming problems and hope that you will find this blog useful for solving your own optimization problems. Optimization problems can be easily solved on SAP Profitability and Performance Management Cloud using Python Remote Function Adapter which offers huge possibilities to end users for different purposes. My name is Novak and in this blog I will share with you the progress that I have made in the field of Linear Programming in SAP Profitability and Performance Management Cloud.
0 kommentar(er)
