Analytic network process (ANP) example

A worked-style explanation of an analytic network process example: how a supermatrix handles feedback loops between interdependent criteria and compares with AHP.

3 min read3 viewsAnalytic network process (ANP) example
— · 0 votes
Contents
  1. Why a network replaces a hierarchy
  2. Building the supermatrix step by step
  3. Reading the final priorities
  4. Questions about the ANP example
  5. What does an analytic network process example actually show?
  6. How does ANP differ from the analytic hierarchy process comparison?
  7. Do I need a supermatrix for interdependent criteria?
  8. Can ANP handle feedback loops in networks of criteria?

An analytic network process (ANP) example shows how to weight criteria when they influence each other, rather than sitting in a clean hierarchy. The method builds a network of interdependent criteria and feedback loops, then places the paired comparisons into a supermatrix to derive final priorities. This page walks through the logic of such an example and explains where it departs from the analytic hierarchy process comparison most people meet first.

Why a network replaces a hierarchy

An analytic network process example begins by drawing criteria as nodes in a network instead of levels in a tree. In that network, one criterion can feed another and receive influence back along a feedback loop, which a strict hierarchy cannot represent. The analytic hierarchy process comparison matters here because AHP assumes criteria are independent and ranked in one direction. ANP keeps the same pairwise questioning but accepts two-way influence between criteria and between clusters.

Building the supermatrix step by step

The supermatrix is the table that stores how every element influences every other element in the network. In a worked example, you compare elements within each cluster, then compare clusters that connect, and each set of comparisons becomes a column of local priorities. Those columns are arranged so that rows and columns match the same nodes, giving an unweighted supermatrix. Weighting the clusters and raising the matrix to powers settles the priorities into a stable set.

Reading the final priorities

The output of an ANP example is a single priority score for each alternative, built from all the influences in the network. You read those scores as relative weights, so an alternative with double the score carries twice the pull across the whole model. Interdependent criteria can lift or lower an option once feedback is counted, which is the effect a hierarchy would miss. Checking the consistency of each comparison set keeps the final priorities defensible.

Questions about the ANP example

What does an analytic network process example actually show?

An analytic network process example shows how to rank options when the criteria affect one another. It draws the criteria as nodes, captures the influence in a supermatrix, and turns the paired comparisons into final priority scores. The example makes the feedback loops visible, so you can see why a plain ranking would give a different answer.

How does ANP differ from the analytic hierarchy process comparison?

ANP differs from the analytic hierarchy process comparison because it allows relationships in both directions between criteria. AHP treats criteria as independent and arranges them in a hierarchy, while ANP treats them as a network with feedback loops. The pairwise comparison step is similar in both, but ANP adds the supermatrix to combine the influences.

Do I need a supermatrix for interdependent criteria?

Yes, interdependent criteria call for a supermatrix because it records the influence each element has on the others. The supermatrix holds local priorities in columns and lets you combine them into global priorities across the network. Without it, the cross-influence between criteria would be lost and the weights would be misleading.

Can ANP handle feedback loops in networks of criteria?

Yes, ANP is built to handle feedback loops in networks of criteria and between clusters. Each loop is entered as an influence in the supermatrix, so a criterion that affects another can also receive influence back. Raising the matrix to powers lets those loops settle into stable priorities instead of circling.

Read next

Latest articles