McdaToolkit 4.0.1-rc3

MCDA Toolkit

MCDA Toolkit is a lightweight .NET library designed to support Multi-Criteria Decision Analysis (MCDA) tasks. It helps developers and analysts define and solve decision-making problems, from simple comparisons to more structured evaluation models.

🚀 Key Features

.NET 6 and .NET Standard 2.0 Support Compatible with both .NET 6 and .NET Standard 2.0, making it suitable for modern applications as well as legacy or cross-platform projects.

⚡ Easy to Use

Designed to be simple and straightforward — no unnecessary setup, no configuration files. You can define and run a decision model with just a few lines of code.

Documentation

<<<<<<< HEAD <<<<<<< HEAD Make sure to read the docs

Make sure to read the docs

c9396c4 (fix(README.md): updated docs path) ======= Make sure to read the docs 99d5ee3 (fix(README.MD): remove /docs/ part from docs path)

Dependencies

• Math.Net Numerics
• Light Results

Quick Example

Here's a simple example to show how to prepare and structure your data for an MCDA calculation using the MCDA Toolkit.

1️⃣ Define the Decision Matrix

Each row represents an alternative, and each column corresponds to a criterion.

double[,] matrix = new double[,]
{
    { 66, 56, 95 },
    { 61, 55, 166 },
    { 65, 49, 113 },
    { 95, 56, 99 },
    { 63, 43, 178 },
    { 74, 59, 140 },
};

2️⃣ Prepare Weights

They must be double values between 0 and 1, and must sum to 1

double[] weights = new double[]
{
    0.4, 0.25, 0.35
};

3️⃣ Define Criteria Types

Each criterion is marked as either a cost (-1) or a benefit (1).

int[] types = new int[]
{
    -1, -1, 1
};

4️⃣ Build the Data Object

var data = DataBuilder
    .Create()
    .AddWeights(weights)
    .AddDecisionCriteria(types)
    .AddDecisionMatrix(matrix)
    .Build();                    

5️⃣ Create MCDA method

var vikor = VikorBuilder
    .Create()
    .WithNormalizationMethod(NormalizationMethod.Vector)
    .WithVParameter(0.5)
    .Build();

6️⃣ Run

var result = vikor.Run(data);

Outcome

The result of the calculation is returned as an object of type Result<Ranking<T>>, provided by the LightResults library. It encapsulates both the Ranking<T> and information about whether the operation succeeded or failed.

Example returned success result:

This approach follows the principles of Railway Oriented Programming, a functional programming pattern that models the flow of data along two possible tracks — one for success and one for failure.

More about Result type can be readed on official LightResult documentation

Example returned success result:

Ranking<double>()
{
    RankingItems = List<RankingRow<double>>()
    {
        RankingRow<double>
        {
            Alternative = 1,
            Rank = 6,
            Score = 0.417
        },
        RankingRow<double>
        {
            Alternative = 2,
            Rank = 2,
            Score = 0.552
        },
        RankingRow<double>
        {
            Alternative = 3,
            Rank = 4,
            Score = 0.54
        },
        RankingRow<double>
        {
            Alternative = 4,
            Rank = 3,
            Score = 0.54
        },
        RankingRow<double>
        {
            Alternative = 5,
            Rank = 5,
            Score = 0.429
        },
        RankingRow<double>
        {
            Alternative = 6,
            Rank = 1,
            Score = 0.568
        }
    }
};