DecisionNetS Suite

November 2007: DecisionNetS Suite of INET modeling webapps is available for demonstration.

What is it?
DecisionNetS Suite webapp tools are browser-based versions of our powerful software application designed to assist people in analyzing complex problems and issues, especially when empirical information is sparse or uncertain. Just like their desktop variants (SIAM and Causeway), the DecisionNetS applications can be used in a range of operational situations, from corporate decision making to national security planning.

As with SAIC's desktop applications, the DecisionNetS webapps help users break down complicated issues into simpler parts, thereby allowing them to recognize and evaluate important relationships among sub-issues more easily. With role-based user interactions for model construction and automated evaluation functions, the DecisionNetS webapps encourage a more detailed level of thinking than is sometimes possible by intuition alone. Once identified, these detailed issues and relationships can be assessed for their importance within the larger picture. To go beyond the traditional desktop applications, the DecisionNetS webapps are accessible with Microsoft's Internet Explorer browser.

The Need for the DecisionNetS Series of Webapps
The stated goal of SAIC’s Influence Net modeling approach is to facilitate investigation of uncertain, diverse, and dynamic situations for which little or no objective data exist. During the years of operational use with our desktop applications, users have created and analyzed Influence Net (INET) models to: advise decision makers of alternate perspectives of current situations; and assist in the assessment of alternative policies and plans to influence future outcomes.
In many instances the team of experts gather together — at the same time and place — to investigate the factors and relationships that underlie a situation. This workshop method has proved beneficial when the decision horizon is extremely short, and human-to-human (physical) interaction is needed to rapidly flesh out consensus. However, the prudent decision maker recognizes that uncertain, dynamic, and diverse situations require an ongoing analysis. For example, long-term examination is required:

Across multiple domains of expertise
Over time as the situation and evidence evolve
As the situation in one region expands to another region, such as international market sales or regional instability

In these cases, modelers and their consumers likely are not co-located, nor available at the same time. Therefore, the need for commonly accessible models, which are editable along multiple lines of reasoning, must be met. For example, suppose you are the decision maker responsible for assessing the feasibility of launching a new product line. Your product development staff — supported by personnel from sales, operations, and research and development (R&D) — help the INET modeling team develop the structure of an INET model. But the preliminary set of INET nodes and links, as well as their parametric attributes, will not remain constant as your company moves from one consumer region to another; nor will a constant INET model reflect the real world as time passes. In such cases, an audit trail of the INET model’s structural evolution is required. Moreover, the INET model must be accessible to remotely located users, who do not have network bandwidth that supports the rich user interaction of a network-deployed desktop application. Finally, the notion of role-based apps is a requirement that is essential to effective collaboration among users of differing interests. It is to address this need for distributed collaboration of INET modelers, analysts, and their consumer that the DecisionNetS series of web applications, including the server-side INET Repository, was developed.

Introduction to SAIC's DecisionNetS Series Webapps
The DecisionNetS Suite webapp tools are designed to facilitate reasoning under uncertainty for role-based user interaction. Specifically, the DecisionNetS suite includes the following webapps:

The DesignS webapp is a browser-accessible version of SAIC's traditional desktop applications, SIAM and Causeway. However, interaction with an server-side model repository allows remotely located users to access models their colleagues have created and completed in part or whole. For example, experts from complementary fields of study are able to "share" their model modifications through the repository that manages the history of the edits, as well as the current state of models. Specifically, DesignS — and DecisionNetS' underlying INET repository — allow users to:

Manage INET models within hierarchy of user-named folders
Create new models with DesignS, or import models from SAIC's desktop applications, as a revision zero version of the INET file
Track revisions of an INET model along the "main" branch of the repository file
Branch off the "main" line of the INET file when structural edits are required to examine an evolving situation, but not to replace earlier drafts of the INET model
Export INET files from the repository for insertion into a second repository — say at another location that does not share network connectivity — or for import into SAIC's desktop applications

This management of model configurations encourages collaboration of colleagues — among themselves, as well as with their consumer, the decision maker. However, in addition to the configuration control provided by the DecisionNetS INET Repository, DesignS includes the following construction and analysis features via the Internet Explorer browser web pages:

Create new INET models to be stored in the repository under a "branch-version" hierarchy
View/Edit INET models previously stored in the repository, to include INET pages, nodes, and links
View/Edit INET nodes and links, to include both textual attributes and parametric values
Add/Edit/Delete excursions within the INET model's hierarchy of parametric assignment
Perform single excursion analyses, including impact and sensitivity analysis for a selected outcome
Perform multiple excursion analyses, including impact and sensitivity analysis for a selected outcome

The DirectS webapp is intended to meet the needs of senior executives and their advisors who want a browser-based method to examine analysis findings of INET models; without working directly with the construction features of DesignS and SAIC's desktop applications. DirectS users can view a previously constructed INET model from their portfolio, select individual factors of interest — as outcomes or influences — and examine the graphical output from a variety of analysis products. DirectS' analysis products include:

Top Influences computes the potential of each factor of interest to change the likelihood of success of the selected goal. The "top 5" influencing factors are diplayed in a "pie chart."
Risk Analysis computes the potential of each factor of interest to DECREASE the likelihood of success of the selected goal.
Gain Analysis computes the potential of each factor of interest to INCREASE the likelihood of success of the selected goal.
Importance Ranking computes the direct impact of factors on the selected goal.
Importance Trend — Bar computes the direct impact of factors on the selected goal for multiple situations. These results are graphically depicted with a 'stacked bar' chart, where the "slices" in each stacked bar illustrate the impact of the corresponding factor on the goal for the associated situation.
Importance Trend — Line computes the direct impact of factors on the selected goal for multiple situations. These results are graphically depicted with a line graph, where each line illustrates the impact of the corresponding factor on the goal across the multiple situations.
Likelihood Analysis computes the likelihood of success for the selected goal across multiple situations.
Multi-Variability Trend Analysis computes the variability in the potential of each factor of interest to change the likelihood of success of the selected goal.
Variability Trend Analysis computes the variability in the potential of a factor of interest to change the likelihood of success of the selected goal.

Although not included as part of the demontration release of our DecisionNetS version 1, the INET development team is working on a third webapp for collaboration among subject matter experts. AdviseS is intended as a browser-based method of gathering domain expertise from a distributed set of users. The expert, or INET modeler, will use AdviseS to create a web-questionnaire accessible by the domain specialist participants. An INET file from the DecisionNetS repository will be selected when creating this questionnaire. The questionnaire then will be configured to have expert-specific questions, along with instructions for participants and categories for expert participation. The expert then will use the AdviseS web application to complete the questionnaire, after which the INET modeler then collects participants responses and merge them onto the INET model for storage in the repository and further analysis with DesignS and/or DirectS. We hope to have a demonstration version of this webapp in 2008.

DecisionNetS Architecture
Readers interested in the underlying architecture of SAIC's DecisionNetS Suite of webapp, are referred to the following figure, which is enlarged by clicking on the image:

Summary
The DecisionNetS Suite of webapps encourages the user to consider and track the impact of all issues, events, perceptions, and other factors which are of interest to the decision maker's goals. By analyzing all relevant issues, the DecisionNetS user can make the most informed decision possible. The automated assessment tools of DesignS DirectS allow the user to evaluate the importance of uncontrolled variables, and adjust decisions to incorporate major developments and changes to the situation or information updates available to the user.

One of the most beneficial uses of the DecisionNetS Suite of webapps is in the area of analytical collaboration. Analysts, decision makers, corporate strategists, and operational planners with expertise in different disciplines can use these webapps to collaborate on issues which require an understanding of many fields. Users who share a specialty can employ the webapps to critique and challenge the logic of one another on any given issue. Changes can then be incorporated and assessed for their impact. And with the cross-excursion and trend analysis tools available with DesignS and DirectS, decision makers are provided with the findings — consensus or divergent — that determine the next step in development of feasible, effective strategic planning.