Help Contents

Overview

Quick Start Guide

Portfolio Lists

Portfolios

Attributes

Prioritization

Optimization

Bubble Charts

Bar Charts

Ranking Charts

Efficient Frontier Charts

Sensitivity Charts

Menus

The Optsee® Pro Optimization Form

Form Overview

Note: For Optsee® Plus, optimizations are set-up in the Optsee® Plus Optimization form.

Use this form in Optsee® Pro to set the parameters for optimizing the selected Portfolio under constraints. You can have up to 100 constraints and essentially unlimited project dependency constraints.

Fields:

These top two fields display the name of the portfolio to be optimized and the name of the parent decision model. You enter the name of the optimization name in the "Name" field. The "No. of Constraints" field displays the number of constraints entered by the user.

Checkbox:

[Save Optimization Results as Portfolio View] checkbox: Checking this checkbox automatically saves the project set selected by the optimization in a Portfolio View.

Optimization Type:

Use these controls set the type of optimizer will be used to optimize the portfolio:

  • [Genetic] radio button: This is the default optimizer in portfolios containing more than 20 projects (1,048,576 possible combinations). This algorithm lets the Optimizer test thousand or millions of possible combinations of projects to determine an optimized or close to optimal set that meets your constraints. The genetic algorithm works by creating an initial set of parent portfolios that meet your constraints, and then combining these parent portfolios in such a way to create a generation of child portfolios. The best combined set of parent and child portfolios are then selected and used to create the next generation of portfolios. This process continues until the user-specified optimization parameters are satisfied and/or the process converges to a single optimized result (i.e., the identical result is obtained after a predetermined number of generations).
  • [Branch and Bound] radio button: This is a proprietary “Branch and Bound” algorithm designed to quickly explore the optimization surface and find an optimal solutions if one exists. It is designed for portfolios containing less than 32 projects. IMPORTANT NOTE: The “Branch and Bound” Optimizer IS DISABLED for finding optimized portfolios using only "Mean not greater" or "Mean not less" constraints as it is not effective for these kinds of optimizations.
  • [Brute Force] radio button: This is the default optimizer in portfolios not more than 20 projects (1,048,576 possible combinations). The "Brute Force" tests all possible portfolio combinations. For portfolios of more than 20 projects, this can take many hours or days (or years(!) for larger portfolios).

Optimization Parameters:

These controls set the parameters that the Optimizer will use to optimize the portfolio.

Optimization Attribute to Maximize popup menu: This popup menu is used to select the attribute that you want to maximize. The default is the SMART Score. When you optimize to maximize SMART Score, the Optimizer ranks the portfolios based on highest total SMART Score. It finds sets of projects that have the highest SMART Score and come closest to meeting your constraint settings. You can also optimize to an attribute other than SMART Score as well as optimizing using efficient frontier ratios as described in the [Use Efficient Frontier] Check Box section below.

When the [Use Efficient Frontier] Check Box is checked, this popup is labeled "Efficient Frontier Numerator (Y-Axis) and the "Efficient Frontier Denominator (X-Axis)" popup menu becomes enabled.

See also Comparing Portfolios Against Efficient Frontiers for more information on using Efficient Frontier charts for comparing portfolios.

Efficient Frontier Denominator (X-Axis) popup menu: This popup menu is enabled after the [Use Efficient Frontier] Check Box is checked.

[Use Efficient Frontier] Check box: You can optimize to an attribute other than SMART Score in two different ways: using a ratio of SMART Score to the optimized attribute or using the attribute value itself.

Optimizing along the Efficient Frontier curve using the attribute ratio will find optimized portfolios that yield the most value per unit of the optimized attribute. For example, consider the two projects in the table below:

 
Profit
Cost
Profit/Cost
Project A
750
250
3
Project B
600
150
4
  • If these two projects were ranked based on maximizing Profit, Project A would be ranked higher than Project B because it has the higher Profit (750).
  • If these two projects were ranked based on maximizing the Profit/Cost ratio, Project B would be ranked higher than Project A because it has the higher Profit-to-Cost ratio. Project B yields four units of Profit for every unit of cost whereas Project A yields three units of Profit for every unit of cost. In other words, Project B is a better value for the cost than Project A.

If the [Use Efficient Frontier] check box is not checked, the Optimizer will optimize to maximize the total value of the selected attribute and will not consider the SMART Score value of the optimized portfolio at all. For example, if you were to optimize to an attribute called "Profit," the Optimizer would find sets of projects that have the highest overall profit and come closest to meeting your constraint settings regardless of the SMART Score of the final portfolio.

Optimizer Parameter Settings

If the [Genetic Optimization] radio button is selected, the following popup menus are displayed:

  • Maximum Number of Parents popup menu: This popup menu is used to select the maximum number of parents that will be created in the initial population. The optimizer will try to make as many parent portfolios as it can up to this number. However, it may not be able to make the maximum amount due to being unable to find enough portfolios that meet the applied constraints.
  • Maximum Number of Generations popup menu: This popup menu is used to select the maximum number of generations (including the parent generation) that will be created in the optimization process. The optimizer keeps creating new generations until this number has been satisfied, or until the specified number of "Repeats Indicating Convergence" has been reached.
  • Repeats Indicating Convergence popup menu: This popup menu is used to select the maximum number of consecutive generations that can be created that have an identical optimized attribute maximum value. Thus, when three or more generations have repeated the same optimized attribute maximum value, the optimization process has converged on the optimum value for that portfolio and the process will cease.
  • Percentage of Mutations popup menu: This popup menu is used to select the percentage of mutations to be added to each generation created after the initial population has been created. Mutations are new portfolios that are created using the same algorithm as the parents, and are added to the population to add diversity in the population and prevent premature convergence before a high optimization is achieved. However, too large a percentage of mutations can prevent the optimization from converging at a high optimization due to dilution of the population with initial population (non-optimized) solutions.

There are no parameters to assign if the [Branch and Bound] or [Brute Force] radio buttons are selected:

Constraint Controls:

These controls set the constraints that the optimizer algorithm will apply when optimizing the portfolio. You can constrain either the sum total attribute values or the average of the sum total (sum total/number of projects).

The constraints are displayed in the Constraints table on the form. The description of the controls for each constraint parameter are as follows:

Select Optimization Constraint popup menu: This popup menu is used to select individual attributes that you want to constrain. The attribute name will be added to the "Attribute Constraint" column of the row selected (highlighted) in the Constraints table of the form. This object is disabled if no row is selected. The attribute order in the table does not affect the optimization results.

Constraint Type check box: This popup menus is used to select individual attributes that you want to constrain. There are 4 types of constraints:

  • Not greater: Final sum-total value in the optimized portfolio will not be greater than constrained value
  • Not less: Final sum-total value in the optimized portfolio will not be less than the constrained value (the optimizer is designed to maximize above this value, not get close to it)
  • Mean not greater: Final mean (average) of the sum total of all the values for that attribute (Sum total/Number of Projects) in the optimized portfolio will not be greater than constrained value.
  • Mean not less: Final mean (average) of the sum total of all the values for that attribute (Sum total/Number of Projects) in the optimized portfolio will not be less than the constrained value.

Maximum, Minimum and Mean Value fields: These display-only fields show the maximum, minimum, and mean values. The minimum value is the minimum value for that attribute in the portfolio, and the Maximum value is the sum total of all the values for that attribute minus the minimum value (the maximum value that can be considered constrained).

Value field: You enter the actual value that you want to use to constrain the attribute selected in the Select Optimization Constraint popup menu. The default is the Minimum Value. This actual value must be equal to or within the values displayed in the Maximum and Minimum Value fields.

[Clear All] Button: Click this button to clear all rows in the Constraints table, and create an empty row at the top.

[Delete Line] Button: Click this button to delete the selected (highlighted) row in the Constraints table. This button is disabled if no rows have been selected.

[Insert Line] Button: Click this button to insert a row below the selected (highlighted) row in the Constraints table. This button is disabled if no rows have been selected. This button is disabled after six rows have been added to the Optimizer.

[Add Line] Button: Click this button to add a row below the selected (highlighted) row in the Constraints table. This button adds a row to the bottom if no rows have been selected. This button is disabled after six rows have been added to the Optimizer. No. of Constraints Field: Displays the number of constraints created (up to 100 maximum).

Dependencies:

You can select a set of dependencies to use as constraints in an optimization by selecting the name of the dependency set from the "Select Dependency Set:" popup menu at the bottom of the list box.

Dependency sets are created using the Portfolio Dependency Set Form and the Portfolio Dependency Editor Form.

Optimizer Form Buttons:

[Suggested Parameters] button: Click on this button to set the form to suggested parameters for the Genetic Optimizer or select a different optimizer based on the number of projects in the portfolio.

[Last Used Parameters] button: Click on this button to restore the last set of values used on this portfolio since it was opened. Note that this button is disabled until one optimization has been performed. The dependency set that is loaded when the [Last Used] is based on the current version of that set, and may be different than the one saved and displayed if the archived version of the set has been modifed since it was last used.

[Optimize] button: Click on this button to run the optimization. You will see an Optimizer Progress form that displays the progress of the optimization. Optimizations can be very computationally intense and can take several minutes to multiple hours to complete depending on the size of the portfolio and the number of the constraints. Once complete, the Optimize Portfolio Form is automatically closed, the Optimization Results Form is displayed, and the optimized portfolio is displayed in the "Last" column of the Portfolio form (if it is open). This button is disabled if the Name field is blank or there are rows in the Constraints table missing values.

[Cancel] button: Click on this button to cancel the optimization and close the form.