flutter_decision_making 1.0.3

Dart package for lightweight decision making using the Analytic Hierarchy Process (AHP). Helps prioritize alternatives objectively via pairwise comparison.

A Flutter package for implementing criteria decision-making using the Analytic Hierarchy Process (AHP).

Easily manage criteria, alternatives, pairwise comparisons, consistency checks, and final decision scoring — all in one unified package.

🧠 This package implements the Analytic Hierarchy Process (AHP) originally developed by Thomas L. Saaty.

---

✨ Features

• Generate hierarchy from criteria and alternatives
• Pairwise comparisons using Saaty's 1–9 scale
• Requires manual filling of paired comparison values for criteria and alternatives for valid analysis
• Complete validation and exceptions if there are values that have not been filled in
• Consistency Ratio check to ensure logical consistency
• Eigenvector and final score calculation
• Customizable and extendable architecture
• Built-in performance profiling (dev-friendly)

---

📚 Usage Guide

Initialize the FlutterDecisionMaking instance before using other methods.

late FlutterDecisionMaking _decisionMaking;

 @override
  void initState() {
    super.initState();
    _decisionMaking = FlutterDecisionMaking();
 }

🛠️ User Role

1. Define Criteria and Alternatives

Each item must have a unique ID. If not provided, the package auto-generates it.

final criteria = [
  Criteria(id: 'c1', name: 'Price'),
  Criteria(id: 'c2', name: 'Quality'),
];

final alternatives = [
  Alternative(id: 'a1', name: 'Product A'),
  Alternative(id: 'a2', name: 'Product B'),
];

2. Identification, Generate Hierarchy and Generate Pairwise Comparison Input

Validate and prepare your inputs.

await _decisionMaking.generateHierarchyAndPairwiseTemplate(
listCriteria: criteria,listAlternative: alternatives);

After the process is complete, a paired matrix list will be generated:

List<PairwiseComparisonInput<Criteria>> inputCriteria;

List<PairwiseAlternativeInput> inputAlternative;

Please ensure that all priority weights are filled in before proceeding to the next step.

3. Input Pairwise Matrix and Generate Result

Saaty scale:

You can use custom descriptions for each value when displaying the scale to users, but the numeric values must still conform to the Saaty scale.

The package only accepts values from 1 to 9 for each comparison.

final List<PairwiseComparisonScale> pairwiseComparisonScales = [
  PairwiseComparisonScale(
    id: '1',
    description: "Equal importance of both elements",
    value: 1,
  ),
  PairwiseComparisonScale(
    id: '2',
    description: "Between equal and slightly more important",
    value: 2,
  ),
  PairwiseComparisonScale(
    id: '3',
    description: "Slightly more important",
    value: 3,
  ),
  PairwiseComparisonScale(
    id: '4',
    description: "Between slightly and moderately more important",
    value: 4,
  ),
  PairwiseComparisonScale(
    id: '5',
    description: "Moderately more important",
    value: 5,
  ),
  PairwiseComparisonScale(
    id: '6',
    description: "Between moderately and strongly more important",
    value: 6,
  ),
  PairwiseComparisonScale(
    id: '7',
    description: "Strongly more important",
    value: 7,
  ),
  PairwiseComparisonScale(
    id: '8',
    description: "Between strongly and extremely more important",
    value: 8,
  ),
  PairwiseComparisonScale(
    id: '9',
    description: "Extremely more important (absolute dominance)",
    value: 9,
  ),
];

Call this method to compute the final scores based on input data.

await _decisionMaking.generateResult()

🛠️ AHP Role

on generate result, AHP will do

1. Generate Pairwise Matrix

Create a pairwise comparison matrix from criteria or alternatives input.

2. Calculate Approximate Eigenvector

Calculate priority weights (eigenvector) from the pairwise matrix.

3. Check Consistency Ratio

Compute and verify matrix consistency using the Consistency Ratio (CR).

4. Calculate Result

Calculate final scores for alternatives based on criteria and alternative weights, including consistency checks.

The result includes:

• Sorted list of alternative scores (results)
• Consistency status and ratio for criteria and alternatives
• Additional notes on consistency issues (if any)

class AhpResult {
  final List<AhpResultDetail> results;
  final bool isConsistentCriteria;
  final double consistencyCriteriaRatio;
  final bool isConsistentAlternative;
  final double consistencyAlternativeRatio;
  final String? note;

  AhpResult({
    required this.results,
    required this.isConsistentCriteria,
    required this.consistencyCriteriaRatio,
    required this.isConsistentAlternative,
    required this.consistencyAlternativeRatio,
    this.note,
  });
}

class AhpResultDetail {
  final String? id;
  final String name;
  final double value;

  AhpResultDetail({
    required this.id,
    required this.name,
    required this.value,
  });
}

---

🔍 Sample Case

1. Pairwise Comparison Matrix

	C1	C2	C3
C1	1.000	1.000	5.000
C2	1.000	1.000	7.000
C3	0.200	0.143	1.000

2. Matrix Normalization

Column totals:

• C1: 2.200
• C2: 2.143
• C3: 13.000

Normalized matrix:

	C1	C2	C3
C1	0.455	0.467	0.385
C2	0.455	0.467	0.538
C3	0.091	0.067	0.077

3. Eigenvector (Priority of Each Criterion)

Criterion	Priority
C1	0.436
C2	0.487
C3	0.078

Interpretation:

• C2 is the most important criterion (48.7%)
• Followed by C1 (43.6%)
• C3 has a low weight (7.8%)

4. Consistency Ratio (CR)

a. λ_max Value

λ_max = 3.078

b. Consistency Index (CI)

CI = (λ_max - n) / (n - 1) = (3.078 - 3) / 2 = 0.039

c. Consistency Ratio (CR)

Random Index (RI) for n=3 is 0.58

CR = CI / RI = 0.039 / 0.58 ≈ 0.067

5. Conclusion

• Consistency Ratio (CR) = 0.067 < 0.1, which means the level of consistency is still acceptable.
• The criterion weights can be used for further decision-making.

---

⚙️ Architecture Notes

• Uses immutable data classes (Criteria, Alternative, etc.)
• Unique ID generation via internal _helper
• Integrated performance profiling using Stopwatch
• Strong validation with helpful exceptions:
  • Duplicate IDs
  • Empty inputs
  • Invalid matrix dimensions

---

📈 Performance Profiling

Every major method logs:

• Start and end timestamps.
• Execution duration (in milliseconds).

Great for debugging and optimization during development.

---

📌 Important

Because the calculations are performed on the client side, the total number of criteria and alternatives may impact your device’s performance. Please use the data wisely.

---

🙋‍♂️ Contributing

We welcome all contributions and suggestions!

👉 Open an issue or submit a pull request at GitHub Repo