Multi-Criteria Decision Analysis
Definition
Multi-Criteria Decision Analysis (MCDA) is a framework for making spatial decisions that balance competing objectives—cost, risk, equity, environmental impact—by scoring and weighting multiple criteria. Steps include criteria selection, normalization to comparable scales, weighting (stakeholder-driven or analytic hierarchy process), and combination (additive, multiplicative, ordered weighted averaging). Sensitivity analysis tests how results change with weights, revealing robust choices. Transparency is essential: document assumptions and trade-offs. MCDA maps are not single truths; they surface options and the value judgments behind them. When criteria involve uncertain or scenario-based rasters, MCDA can integrate ranges rather than single numbers. Provide explicit methods, QA notes, and version history so others can reuse the layer responsibly. Provide explicit methods, QA notes, and version history so others can reuse the layer responsibly. Provide clear usage notes, QA artifacts, and version history to aid reuse and review. Provide clear usage notes, QA artifacts, and version history to aid reuse and review. Provide clear usage notes, QA artifacts, and version history to aid reuse and review. Provide clear usage notes, QA artifacts, and version history to aid reuse and review.
Application
Site selection for facilities (hospitals, warehouses, wind farms), route planning that balances cost and impact, conservation prioritization, and disaster mitigation investments commonly use MCDA. Participatory processes engage communities to co-create weights and constraints so outcomes are legitimate and actionable.
FAQ
How are weights determined fairly?
Use structured methods (AHP, pairwise comparisons) with diverse stakeholders, publish results, and explore alternative weighting schemes.
Can constraints be hard rather than soft?
Yes—mask out no-go areas (protected zones, setbacks) before scoring. MCDA then ranks feasible sites only.
How to avoid ‘garbage in, garbage out’?
Validate inputs, normalize consistently, and run sensitivity tests. Provide uncertainty maps and avoid overstating precision.
What if criteria conflict strongly?
Present Pareto frontiers—sets of options where improving one criterion worsens another—so decision makers choose explicitly among trade-offs.