Knowledge · Decision Analysis

AHP and ANP software: analysing complex decisions transparently

AHP and ANP structure multi-criteria decisions, make judgements traceable and combine qualitative assessments with mathematically derived priorities. This page explains the core logic – from hierarchies and pairwise comparisons to consistency, supermatrices, ratings and sensitivity analysis.

Starting point

Why use AHP and ANP software?

Many real decisions are multi-criteria problems: several objectives, qualitative and quantitative criteria and multiple alternatives must be considered at the same time. A simple score can hide how weights were derived and how stable the result is when assumptions change.

AHP and ANP software therefore supports more than ranking. It helps make the decision structure visible, capture judgements systematically, detect inconsistencies and investigate the robustness of the result.

Analytic Hierarchy Process

What is AHP?

The Analytic Hierarchy Process (AHP), developed by Thomas L. Saaty, is a method for structured analysis of multi-criteria decisions. A decision problem is organised as a hierarchy: an overall goal is broken down into criteria and, where appropriate, sub-criteria; the alternatives sit at the lower level.

Weights are not simply assigned arbitrarily. AHP uses pairwise comparisons. Two elements are compared at a time with respect to their importance or preference relative to a parent element. Mathematical procedures convert these judgements into local priorities and then overall priorities.

Define the decision model

Specify the goal, criteria, sub-criteria and alternatives.

Represent the problem as a hierarchy

Arrange the elements so that their role in the decision model is clear.

Perform pairwise comparisons

Assess criteria and alternatives relative to one another.

Synthesise priorities

Combine local weights into overall priorities for the alternatives.

Test sensitivity

Vary selected weights to see how robust the ranking is.

Pairwise comparisons

How are relative priorities determined?

Pairwise comparisons reduce a complex weighting task to manageable judgements: How important is element A compared with element B? The Saaty scale is commonly used.

ValueBasic meaning
1equal importance or preference
3slight to moderate preference
5strong preference
7very strong preference
9extreme preference
2, 4, 6, 8intermediate judgements

Reciprocal values are used for the reverse direction of comparison. The completed comparison matrix is then used to derive relative priorities.

Quality of judgements

Why is consistency analysis important?

Pairwise judgements can be contradictory. If A is strongly preferred to B and B is strongly preferred to C, the assessment of A relative to C should broadly fit these judgements. AHP therefore includes a mathematical consistency analysis.

The maximum eigenvalue is used to derive a consistency index and the Consistency Ratio (CR). It indicates how far the judgement matrix departs from perfect consistency relative to random judgements. A value around 0.10 or lower is often used as a rule of thumb, while the substantive interpretation should always reflect the decision context.

Important: a formally calculated ranking is not automatically a good decision. Transparent models should also show whether the underlying judgements are sufficiently consistent and substantively plausible.

Analytic Network Process

How does ANP differ from AHP?

The Analytic Network Process (ANP) extends the hierarchical logic of AHP to networks with dependencies and feedback. In real decisions, criteria, alternatives or other elements may depend on or influence each other. A strict hierarchy can represent these relationships only to a limited extent.

AHP

A hierarchical structure with goal, criteria, optional sub-criteria and alternatives. Particularly useful when the problem can be represented meaningfully as levels.

ANP

A network structure with clusters, nodes, interdependencies and feedback. Particularly useful when elements influence one another and those dependencies need to be modelled explicitly.

AHPANP
HierarchyNetwork
primarily directed level structuredependencies and feedback can be represented
synthesis through the hierarchysynthesis through supermatrices
well suited to clearly structured criteria modelswell suited to interdependent, networked decision models

Mathematical core of ANP

What is a supermatrix?

The supermatrix brings together the local priorities of connected elements or clusters in one matrix. Where no influence relationship exists, the corresponding matrix blocks carry no influence; where a relationship exists, the local priorities derived from the relevant pairwise comparisons are entered.

Three stages are commonly distinguished:

Unweighted supermatrix

Contains the local priorities of the modelled influence relationships.

Weighted supermatrix

Also reflects the relative importance of clusters so that the matrix is suitable for iteration.

Limit matrix

Obtained through repeated powering and represents the long-run stabilised priorities of the network model.

Result

The resulting priorities show the importance of the elements while taking the modelled interdependencies into account.

Robustness

What does sensitivity analysis show?

A ranking alone does not show how stable a decision is. Sensitivity analysis systematically varies selected weights and observes the effect on the priorities of the alternatives.

If an alternative remains preferred across a broad range of weights, the result is relatively robust. If the ranking changes after small variations, the result depends more strongly on the underlying assumptions. Good decision-analysis software makes such thresholds and rank changes visible.

Efficient evaluation of many alternatives

When is a ratings approach useful?

Complete pairwise comparison becomes cumbersome when a large number of alternatives must be evaluated. The number of required comparisons grows rapidly with the number of objects.

A ratings approach addresses this by first defining a calibrated scale for each relevant criterion. New or additional alternatives are then evaluated individually against those scales rather than compared pairwise with every other object.

This is useful for larger sets of projects, product ideas, technology options, suppliers, investment alternatives or other objects to be evaluated.

Applications

Where are AHP and ANP used?

AHP and ANP are suitable for multi-criteria decisions where quantitative and qualitative criteria need to be considered together.

Strategy & investment

Evaluate strategic alternatives, investments and business options systematically.

Technology & innovation

Compare technology options, product ideas and innovation projects across multiple criteria.

Projects & portfolios

Assess project priorities, portfolios and constrained resources transparently.

Procurement & suppliers

Consider price, quality, risk, capability and other criteria together.

Products & locations

Analyse variants, locations and market options using transparent criteria models.

Sustainability & transformation

Combine economic, environmental, social and strategic criteria within a common decision structure.

Principle example: a company evaluates several technology options using economics, strategic fit, implementation risk and sustainability. If the criteria are largely independent, AHP can provide a clear hierarchy. If technology, costs, risks and strategic options influence one another, an ANP network can represent the real structure more fully.

Software requirements

What should professional AHP/ANP software provide?

As model complexity increases, practical implementation becomes demanding without appropriate software. Professional AHP/ANP software should therefore support the entire decision process transparently rather than simply calculate a ranking.

  • model hierarchies and networks
  • capture pairwise-comparison matrices systematically
  • support the Saaty scale and reciprocal judgements
  • calculate local and global priorities
  • check consistency automatically
  • calculate and document ANP supermatrices
  • support sensitivity analysis
  • support ratings models for larger sets of alternatives
  • visualise model structure and results
  • document and export calculations and results transparently

Putting AHP and ANP into practice

LPS ANP Analyzer is a local Windows application for modelling and analysing hierarchical AHP and network-based ANP decision models. Pairwise comparisons, consistency analysis, ratings, supermatrices, sensitivity analysis and professional result documentation are combined in one workflow.

Explore LPS ANP Analyzer

Frequently asked questions

AHP and ANP in brief.

What is AHP software?

AHP software supports hierarchical multi-criteria decision modelling, pairwise comparisons, priority calculation, consistency analysis and result evaluation.

What is ANP software?

ANP software extends AHP logic to networks, dependencies and feedback, typically synthesising the model through supermatrices.

What is the difference between AHP and ANP?

AHP primarily uses a hierarchy, while ANP can explicitly represent interdependencies and feedback among elements and clusters.

What is a pairwise comparison?

Two elements are compared at a time with respect to relative importance or preference. Multiple judgements form a comparison matrix.

What does the Consistency Ratio mean?

The Consistency Ratio indicates how consistent pairwise judgements are relative to random judgements and serves as a diagnostic for judgement quality.

What is a supermatrix in ANP?

A supermatrix combines local influence priorities in a network model. Weighting and iteration lead to the limit matrix with stabilised priorities.

When are ratings preferable to full pairwise comparison?

For many or regularly added alternatives, a calibrated ratings system can be much more efficient than comparing every object pairwise with every other object.

Further reading

Saaty, T. L. (1980): The Analytic Hierarchy Process. McGraw-Hill. · Saaty, T. L. (1996): Decision Making with Dependence and Feedback: The Analytic Network Process. RWS Publications.