Hi Everyone,
Henri (@henri_wajsblat), Tristan (@tristan.colgate) and I are starting a new series of discussion topics. This week's topic will be about Anaplan's Optimizer feature.
Optimizer is a relatively new feature that uses linear programming to enable end users to solve complex problems by maximizing or minimizing an objective within specified constraints. When I first read about it, I was immediately reminded of Excel's Solver add-on, and I do think the two are pretty similar. However, Anaplan's Optimizer can work on much larger datasets and has security based on the access to the dashboard. Anaplan's Optimizer is also powered by Gurobi, which is the fastest optimization solver. Optimizer has many use cases, many of which can be found on Anaplan'swhite paper. Optimizer can be used for "what-if" scenarios and help organizations plan based on the best option.
There are a few steps to take before setting up Optimizer:
- Identify your Objective(for Linear Programming)
- Helps guide the Optimizer to a solution
- 可以be a maximum or minimum value, such as maximum revenue or minimum cost
- 可以be generated by a formula (with limit on the number of levels of indirection)
- Not needed for Feasibility Programming
- Define Variable(s)
- What value(s) will the Optimizer output?
- 可以be integer (whole number), real (decimal number), or binary (0 or 1)
- Define Constraint(s)(optional)
- limit on the value (max, min, no negative numbers).
- Needs to be Boolean (True/False) with summary of ALL
- 可以be generated by a formula (with limit on the number of levels of indirection)
Some problems may not have a feasible or optimal solution--and an error will be output in those cases. There may also be limitations on performance, the number of decimals, and the formulas used for the objectives and constraints. I would suggest reading the Anapedia sections onLimitationsandIndirectionto understand more about these limitations. Lastly, Optimizer can only be used to solvelinear problems.
The actual setup of Optimizer is pretty simple. To create the Optimizer action, click onActionsin the Model Settings. From theNew Actionmenu, selectOptimizer.
From there, you can name the action(Button Text),set theTime Out(number of seconds after which the Optimizer will stop processing), select whether to useLinear Programming(selecting the optimal solution) orFeasibility Programming(selecting a possible solution in the feasible region). An example of the feasible region can be illustratedhere. If you selectLinear Programming, you will need to select yourObjective. In bothLinearandFeasibility Programming,you will need to enter theVariable(s).Constraint(s)are optional for both.
To try out Optimizer and to get an understanding of how the formulas for the Objective and Constraints can be created, I highly recommend completing thetraining.
If you want to learn more about optimizer, are interested in a demo, want to share how your organization uses optimizer, or have any discussion topic suggestions, please leave a comment below to let us know!
Additional Training and Resources:
Anapedia: Optimizer
Optimizer: Under the Hood
Optimizer: Exercise Walkthrough
Optimizer: Dynamic Time Setup
- Diana Xu
