optimatika / okAlgo

Idiomatic Kotlin extensions for ojAlgo

okAlgo

Idiomatic Kotlin extensions for ojAlgo, with some inspirations from PuLP.

Linear Algebra DSL

Below is an example of how to use the linear algebra DSL. In this particular example, we create a Markov chain to calculate the probability of 5 consecutive heads in 10 coin flips.

import org.ojalgo.okalgo.populate
import org.ojalgo.okalgo.primitivematrix
import org.ojalgo.okalgo.times

fun main() {
	
	val transitionMatrix = primitivematrix(rows = 6, cols = 6) {
		populate {row, col ->
			when {
				col == 0L -> .50
				row + 1L == col -> .50
				row == 5L && col == 5L -> 1.0
				else -> 0.0
			}
		}
	}

	println("\r\nTransition Matrix:")
	println(transitionMatrix)

	val toTenthPower = generateSequence(transitionMatrix) { it * transitionMatrix }.take(10).last()
	println("\r\nTransition Matrix Raised to 10th Power")
	println(toTenthPower)

	println("\r\nMARKOV CHAIN RESULT: ${toTenthPower[0,5]}")
}

// REFERENCE: https://www.quora.com/What-is-the-probability-of-getting-5-consecutive-heads-in-10-tosses-of-a-fair-coin

OUTPUT:

Transition Matrix:
org.ojalgo.matrix.PrimitiveMatrix < 6 x 6 >
{ { 0.5,	0.5,	0.0,	0.0,	0.0,	0.0 },
{ 0.5,	0.0,	0.5,	0.0,	0.0,	0.0 },
{ 0.5,	0.0,	0.0,	0.5,	0.0,	0.0 },
{ 0.5,	0.0,	0.0,	0.0,	0.5,	0.0 },
{ 0.5,	0.0,	0.0,	0.0,	0.0,	0.5 },
{ 0.5,	0.0,	0.0,	0.0,	0.0,	1.0 } }

Transition Matrix Raised to 10th Power
org.ojalgo.matrix.PrimitiveMatrix < 6 x 6 >
{ { 0.5546875,	0.267578125,	0.1298828125,	0.0634765625,	0.03125,	0.109375 },
{ 0.6015625,	0.287109375,	0.1376953125,	0.06640625,	0.0322265625,	0.140625 },
{ 0.7109375,	0.333984375,	0.1572265625,	0.07421875,	0.03515625,	0.2041015625 },
{ 0.9609375,	0.443359375,	0.2041015625,	0.09375,	0.04296875,	0.333984375 },
{ 1.5244140625,	0.693359375,	0.3134765625,	0.140625,	0.0625,	0.6015625 },
{ 2.78125,	1.2568359375,	0.5634765625,	0.25,	0.109375,	1.15625 } }

MARKOV CHAIN RESULT: 0.109375

MIP Solver DSL

EXAMPLE 1

expressionsbasedmodel {

    val v1 = variable(lower = 3, upper = 6)
    val v2 = variable(lower = 10, upper = 12)

    expression(weight = 1) {
        set(v1, 1)
        set(v2, 1)
    }

    maximise()

    println("v1=${v1.value.toDouble()} v2=${v2.value.toDouble()}")
}

EXAMPLE 2

val model = ExpressionsBasedModel()
        
val v1 = model.variable(lower = 3, upper = 6)
val v2 = model.variable(lower = 10, upper = 12)

model.expression(weight=1) {
    set(v1, 1)
    set(v2, 1)
}

model.maximise()

println("v1=${v1.value.toDouble()} v2=${v2.value.toDouble()}")

Expression building with Kotlin extensions is also being explored:

EXAMPLE 3

expressionsbasedmodel {

    val v1 = variable(lower = 2, upper = 10, isInteger = true)
    val v2 = variable(lower = 2, upper = 10, isInteger = true)

    expression(v1 + 2*v2) {
        weight(1)
    }

    expression {
        set(v1 + v2 EQ 16)
    }

    minimise().run(::println)

    println("v1=${v1.value.toDouble()} v2=${v2.value.toDouble()}")
}

Artifact Instructions

Until this gets deployed to Maven Central, you can use JitPack to import this project as a dependency.

Maven

<dependency>
    <groupId>org.ojalgo</groupId>
    <artifactId>okalgo</artifactId>
    <version>0.0.2</version>
</dependency>

Gradle

compile 'org.ojalgo:okalgo:0.0.2'