implement wpm method

Author: jbytecode

src/main/scala/wpm.scala

@@ -0,0 +1,43 @@
+package org.expr.mcdm
+
+case class WpmResult(
+    normalizedDecisionMatrix: Mat,
+    scores: Vec,
+    ordering: VecInt,
+    bestIndex: Int
+) extends MCDMResult
+
+def wpm(
+    mat: Mat,
+    weights: Vec,
+    directions: Array[Direction],
+    normalization: NormalizationFunction =
+      Normalization.DivideByColumnMaxMinNormalization
+): WpmResult =
+
+  val (n, p) = Matrix.size(mat)
+
+  val normalizedDecisionMat = normalization(mat, weights, directions)
+
+  /*
+  var scoreMat = Array.tabulate(n, p)((i, j) => 0.0)
+  for (i <- 0 until p) {
+    for (j <- 0 until n) {
+      scoreMat(j)(i) = math.pow(normalizedDecisionMat(j)(i), weights(i))
+    }
+  }
+  */  
+  val scoreMat = Array.tabulate(n, p)((i, j) => math.pow(normalizedDecisionMat(i)(j), weights(j)))
+
+  val scores = scoreMat.map(_.product)
+
+  val orderings = scores.zipWithIndex.sortBy(_._1).map(_._2)
+
+  val bestIndex = orderings.last
+
+  WpmResult(
+    normalizedDecisionMat,
+    scores,
+    orderings,
+    bestIndex
+  )

src/test/scala/testwpm.scala

@@ -0,0 +1,42 @@
+import munit.Assertions as A
+
+import org.expr.mcdm.wpm
+import org.expr.mcdm.Direction.{Maximize, Minimize}
+import org.expr.mcdm.Matrix
+
+class TestWpm extends munit.FunSuite {
+  test("Wpm Example - 1") {
+    val decmat = Array(
+      Array(3.0, 12.5, 2, 120, 14, 3),
+      Array(5.0, 15, 3, 110, 38, 4),
+      Array(3.0, 13, 2, 120, 19, 3),
+      Array(4.0, 14, 2, 100, 31, 4),
+      Array(3.0, 15, 1.5, 125, 40, 4)
+    )
+
+    val weights = Array(0.221, 0.159, 0.175, 0.127, 0.117, 0.201)
+
+    val directions =
+      Array(Maximize, Minimize, Minimize, Maximize, Minimize, Maximize)
+
+    val result = wpm(decmat, weights, directions)
+
+    val expectedScores = Array(0.7975224331331, 0.7532541470585,
+      0.7647463553356, 0.7873956894791, 0.7674278741782)
+
+    val expectedBestIndex = 0
+
+    val expectedOrdering = Array(1, 2, 4, 3, 0)
+
+    A.assert(
+      Matrix.elementwise_equal(result.scores, expectedScores, 1e-5)
+    )
+
+    A.assertEquals(result.bestIndex, expectedBestIndex)
+
+    expectedOrdering.zip(result.ordering).foreach { case (a, b) =>
+      A.assertEquals(a, b)
+    }
+
+  }
+}