The Analysis stage shows the numerical story of your model. Results use the limit-matrix settings stored on the current network (Structure → Inspector).

You can also press F5 or use Compute → Show Analysis.

Synthesis (main report)

Left nav under Synthesis scrolls the HTML report:

  1. Unscaled Supermatrix — local priorities by column
  2. Cluster matrix — cluster-to-cluster weights
  3. Scaled Supermatrix — weighted, column-stochastic matrix
  4. Limit Matrix — long-run matrix, with a summary of the method and options used
  5. Global priorities — node priorities from the limit matrix
  6. Subnetwork synthesis results — when nested models exist
  7. Alternative Scores — priorities for alternatives (after synthesis if needed)

Open a sample (e.g. 02_ahp_best_car.anpstudio), enter Analysis, and walk that list top to bottom.

Limit matrix settings

On Structure, select the network and set Limit matrix in the Inspector:

  • Calculus — default SuperDecisions / pyanp-style calculus limit
  • New Hierarchy — hierarchy/network decomposition; optional With limit
  • Sinks — sinks decomposition; Straight normalizer on or off

Analysis, sensitivity, and influence use these network defaults. The Researcher stage can still override method on a single command if you need an experiment.

Sensitivity

Under Sensitivity:

  • Interactive — pick a wrt node and slide parameter p; see alternative bars update (row sensitivity).
  • Global — sweep p and plot each alternative’s score as a line.

Influence

Under Influence analysis, tables summarize how adjusting rows moves alternative scores (raw, rank, marginal, total). Use these after the base synthesis looks sensible.

Perspective

Under Perspective analysis, each column is the alternative-score vector obtained by taking ANP row sensitivity toward the limit p → 1 for that node (evaluated near 1, never at exactly 1). Rows are alternatives; columns are nodes.

Researcher (optional)

Compute → Show Researcher opens a notebook-style command panel for inspection (limit, globals, matrices, loads of other models, and more). Type help there for commands. Prefer Analysis for everyday reading of results.

Next: Glossary · User guide home