implement copras

Author: jbytecode

src/main/scala/copras.scala

@@ -0,0 +1,64 @@
+package org.expr.mcdm
+
+case class CoprasResult(
+    normalizedMat: Mat,
+    weightedNormalizedMat: Mat,
+    splus: Vec,
+    sminus: Vec,
+    Q: Vec,
+    Z: Double,
+    maxQ: Double,
+    rankings: VecInt,
+    bestIndex: Int,
+    scores: Vec,
+) extends MCDMResult
+
+def copras(
+    decmat: Mat,
+    weights: Vec,
+    directions: Array[Direction],
+    normalization: NormalizationFunction =
+      Normalization.DivideByColumnnsSumNormalization,
+    options: Map[String, Any] = Map.empty 
+): CoprasResult = 
+
+    val (nrows, ncols) = Matrix.size(decmat)
+
+    val normalizedMat = normalization(decmat, weights, directions)
+    val weightedNormalizedMat = Matrix.weightizeColumns(normalizedMat, weights)
+
+    var sPlus = Matrix.zeros(nrows)
+    var sMinus = Matrix.zeros(nrows)
+
+    for row <- 0 until nrows do
+        for col <- 0 until ncols do
+            if directions(col) == Direction.Maximize then
+                sPlus(row) = sPlus(row) + Matrix.elementat(weightedNormalizedMat, row, col)
+            else if directions(col) == Direction.Minimize then
+                sMinus(row) = sMinus(row) + Matrix.elementat(weightedNormalizedMat, row, col)
+
+
+    var Q = Matrix.zeros(nrows)
+    var Z = sMinus.map(x => 1.0 / x).sum
+
+    for row <- 0 until nrows do
+        Q(row) = sPlus(row) + (sMinus.sum / (sMinus(row) * Z))
+
+    val maxQ = Q.max
+    val scores = Q.map(x => x / maxQ)
+
+    val rankings = scores.zipWithIndex.sortBy(-_._1).map(_._2)
+    val bestIndex = rankings.head
+
+    CoprasResult(
+        normalizedMat,
+        weightedNormalizedMat,
+        sPlus,
+        sMinus,
+        Q,
+        Z,
+        maxQ,
+        rankings,
+        bestIndex,
+        scores
+    )

src/test/scala/testcopras.scala

@@ -0,0 +1,68 @@
+import munit.Assertions as A
+
+import org.expr.mcdm.copras
+import org.expr.mcdm.Direction.{Maximize, Minimize}
+import org.expr.mcdm.Matrix
+
+class TestCopras extends munit.FunSuite {
+  test("Copras Example - 1") {
+
+    val decmat = Array(
+      Array(2.50, 240, 57, 45, 1.10, 0.333333),
+      Array(2.50, 285, 60, 75, 4.00, 0.428571),
+      Array(4.50, 320, 100, 65, 7.50, 1.111111),
+      Array(4.50, 365, 100, 90, 7.50, 1.111111),
+      Array(5.00, 400, 100, 90, 11.00, 1.111111),
+      Array(2.50, 225, 60, 45, 1.10, 0.333333),
+      Array(2.50, 270, 57, 60, 4.00, 0.428571),
+      Array(4.50, 330, 100, 70, 7.50, 1.111111),
+      Array(4.50, 365, 100, 80, 7.50, 1.111111),
+      Array(5.00, 380, 110, 65, 8.00, 1.111111),
+      Array(2.50, 285, 65, 80, 4.00, 0.400000),
+      Array(4.00, 280, 75, 65, 4.00, 0.400000),
+      Array(4.50, 365, 102, 95, 7.50, 1.111111),
+      Array(4.50, 400, 102, 95, 7.50, 1.111111),
+      Array(6.00, 450, 110, 95, 11.00, 1.176471),
+      Array(6.00, 510, 110, 105, 11.00, 1.176471),
+      Array(6.00, 330, 140, 110, 18.50, 1.395349),
+      Array(2.50, 240, 65, 80, 4.00, 0.400000),
+      Array(4.00, 280, 75, 75, 4.00, 0.400000),
+      Array(4.50, 355, 102, 95, 7.50, 1.111111),
+      Array(4.50, 385, 102, 90, 7.50, 1.111111),
+      Array(5.00, 385, 114, 95, 7.50, 1.000000),
+      Array(6.00, 400, 110, 90, 11.00, 1.000000),
+      Array(6.00, 480, 110, 95, 15.00, 1.000000),
+      Array(6.00, 440, 140, 100, 18.50, 1.200000),
+      Array(6.00, 500, 140, 100, 18.50, 1.200000),
+      Array(5.00, 450, 125, 100, 15.00, 1.714286),
+      Array(6.00, 500, 150, 125, 18.50, 1.714286),
+      Array(6.00, 515, 180, 140, 22.00, 2.307692),
+      Array(7.00, 550, 200, 150, 30.00, 2.307692),
+      Array(6.00, 500, 180, 140, 15.00, 2.307692),
+      Array(6.00, 500, 180, 140, 18.50, 2.307692),
+      Array(6.00, 500, 180, 140, 22.00, 2.307692),
+      Array(7.00, 500, 180, 140, 30.00, 2.307692),
+      Array(7.00, 500, 200, 140, 37.00, 2.307692),
+      Array(7.00, 500, 200, 140, 45.00, 2.307692),
+      Array(7.00, 500, 200, 140, 55.00, 2.307692),
+      Array(7.00, 500, 200, 140, 75.00, 2.307692)
+    )
+
+    val weights = Array(0.1667, 0.1667, 0.1667, 0.1667, 0.1667, 0.1667)
+
+    val fns = Array(Maximize, Maximize, Maximize, Maximize, Maximize, Minimize)
+
+    val result = copras(decmat, weights, fns)
+
+    val expectedScores = Array(0.44194, 0.44395, 0.41042, 0.44403, 0.48177,
+      0.44074, 0.42430, 0.41737, 0.43474, 0.44382, 0.46625, 0.48602, 0.45019,
+      0.45825, 0.51953, 0.54265, 0.56134, 0.45588, 0.49532, 0.44788, 0.45014,
+      0.48126, 0.51586, 0.56243, 0.58709, 0.60091, 0.51850, 0.61085, 0.65888,
+      0.75650, 0.61430, 0.63486, 0.65542, 0.72065, 0.77680, 0.82379, 0.88253,
+      1.00000);
+
+    A.assert(Matrix.elementwise_equal(result.scores, expectedScores, 1e-5))
+
+    A.assert(result.bestIndex == 37)
+  }
+}