CodePlexProject Hosting for Open Source Software

The first edition architecture followed the implementation flow chart presented by Rubin
*et al.* [2010]. The flow chart therein detailed three modules - one for each of the core components of Bayesian analysis. The software was developed to aid with problem formulation and prior probability distributions in Block 1, to
calculate likelihood function estimates in Block 2, and to determine posterior probability distributions of model parameters in Block 3. Many of the underlying components of the first edition of the software have been retained by the latest edition.
However, the original architecture has been revised for two main reasons. First, the generality and flexibility of the interfaces needed to be increased, which will help the MAD software support Bayesian model inversion in different scientific
areas in the future. Second, the implementation is now a single module that handles problem formulation, Bayesian analysis, and all computational configuration steps, which means that the MAD software fully automates model inversion and will permit future
parallelization of the process.

- Draft of Specifications & Software Ideas: Block1 PreProcessing Design 5-25-11.doc
- Flow Chart & Software Diagrams: Pre Processing Block Algorithm.pptx
- PPT Connecting Pre-Processing Steps, Expected GUI Functionality, & R Script Examples: Flow chart and Project List.ppt
- Example Toy Problem & Discussion: Block1 Ancillary Example.docx
- R Script: ToyProblemBlock1.R
- GeoStat Modeling R Script: GeoStatModelExamples.R

**Block 2/3 Material**:

- Draft of Specifications & Software Ideas:Block 2 Likelihood Design 5-25-11.doc
- R Script: ToyProblemBlock2.R
- R Script: ToyProblemBlock3.R (Includes posterior weighted re-sampling of anchor #1)
- R Script: ToyProblemBlock3expanded.R (Includes alternate examples of possible posterior visualizations) - Disregard steps 1 & 2.
- Mock-up of Posterior Analysis Tab
- Proposed z vector construction tab
- Block 3 R script for AGU. Only includes posterior anchor distributions currently.
- Likelihood Computation Tab Diagram

**Functionality Diagram for Blocks I,II, and III**

- Please see attached ppt file (tasks and Milestones.ppt) slide 19 for functionality for Block III
- GUI diagram and accompanying notes for function implementation and potential challenges: Diagram,Data Extraction Notes, Data Manipulation, Likelihood Analysis, & Posterior Prediction Notes
- Road map
- Crib notes for file contents and where data is used in Toy Problem GUI.

**Architecture:**

- Sequencing and automation of model inversion with the MAD software

**Drivers:**

**Meeting Notes:**

- Slides from 2/16/12 HPC options and domain scaling
- Slides from 2/2/12 Developing the linkage to MODFLOW
- Meeting notes 3/8/12 Programming and Developmental Tasks

**Toy Problem 2:**

- Formulation, motivation, and likelihood analysis plans. Coordinate locations for anchors and both measurement types.

**Mad Interface:**

- MadInterface allows incorporating new driver in the MAD plugin (MadInterface help file)

Last edited Nov 8, 2013 at 9:19 PM by matthewoverUCB, version 7