TimeSeriesDecrypter/EvolutionaryAlgorithm.swift at main · anantashahane/TimeSeriesDecrypter · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
//
//  EvolutionaryAlgorithm.swift
//  TimeSeriesDecrypter
//
//  Created by Ananta Shahane on 08/10/2023.
//

import Foundation

struct Individual {
    var weightage : [Int: Double]
    var mutationAmplitude : [Int: Double]
    var mutationDirection : [Int : Double]
    var error : Double?

    init(numberofSignals: Int) {
        weightage = [Int: Double]()
        mutationAmplitude = [Int: Double]()
        mutationDirection = [Int : Double]()
        for i in 1...numberofSignals {
            weightage[i] = abs(Double.NormalRandom(mu: 10, sigma: 10))
            mutationAmplitude[i] = 1
            mutationDirection[i] = 0
        }
    }
}

class EvolutionaryAlgorithm {
    let data : [Double]
    var baseTimeSeries : [Int : [Double]]
    let subtimeSeriesCount : Int
    let offspringPopulationSize : Int
    let parentPopulationSize : Int
    let optimal : [Int : Double]
    //Variables
    var bestError = Double.infinity
    var parentPopulation = [Individual]()
    var offspringPopulation = [Individual]()
    var recombinedEntity : Individual
    var progression = [[Int : Double]]()
    var errorProgression = [Double]()
    init(parentPopulationSize : Int, offspringPopulationSize : Int, numberOfWavelets : Int) {
        baseTimeSeries = [:]
        var weightage = [Int : Double]()
        for i in 1...numberOfWavelets {
            baseTimeSeries[i] = GenerateData(amplitude: Double.random(in: 1...10), frequency: Double.random(in: 1...20), phase: Double.random(in: 1...10_000))
            weightage[i] = Double.random(in: 1...10000)
        }
        optimal = weightage
        print("Original weightage: \(weightage)")
        data = MixTimeSeries(timeSeries: baseTimeSeries, weightage: weightage)
        recombinedEntity = Individual(numberofSignals: baseTimeSeries.count)
        subtimeSeriesCount = 3
        self.parentPopulationSize = parentPopulationSize
        self.offspringPopulationSize = offspringPopulationSize
    }

    func Initialise() {
        for _ in 0..<parentPopulationSize {
            let individual = Individual(numberofSignals: baseTimeSeries.count)
            parentPopulation.append(individual)
        }
    }

    func Evaluate(evaluateParentPopulation : Bool) {
        if evaluateParentPopulation {
            for (parentNumber, parent) in parentPopulation.enumerated() {
                let timeSeries = MixTimeSeries(timeSeries: baseTimeSeries, weightage: parent.weightage)
                var error : Double = 0
                for (index, _) in timeSeries.enumerated() {
                    error += abs(timeSeries[index] - data[index])
                }
                parentPopulation[parentNumber].error = error / abs(data.reduce(0, +))
            }
        } else {
            for (offspringNumber, offspring) in offspringPopulation.enumerated() {
                let timeSeries = MixTimeSeries(timeSeries: baseTimeSeries, weightage: offspring.weightage)
                var error : Double = 0
                for (index, _) in timeSeries.enumerated() {
                    error += abs(timeSeries[index] - data[index])
                }
                offspringPopulation[offspringNumber].error = error / abs(data.reduce(0, +))
            }
            let timeSeries = MixTimeSeries(timeSeries: baseTimeSeries, weightage: recombinedEntity.weightage)
            var error = 0.0
            for(index, _) in timeSeries.enumerated() {
                error += abs(timeSeries[index] - data[index])
            }
            recombinedEntity.error = error / abs(timeSeries.reduce(0, +))
        }
    }

    func Recombine() {
        for key in parentPopulation[0].weightage.keys {
            recombinedEntity.weightage[key] = parentPopulation.map({$0.weightage[key] ?? 1}).reduce(0, +) / Double(parentPopulation.count)
            recombinedEntity.mutationDirection[key] = parentPopulation.map({$0.mutationDirection[key] ?? 1}).reduce(0, +) / Double(parentPopulation.count)
            recombinedEntity.mutationAmplitude[key] = parentPopulation.map({$0.mutationAmplitude[key] ?? 1}).reduce(0, +) / Double(parentPopulation.count)
        }
    }

    func Mutation() -> Individual {
        var returnIndividual = recombinedEntity
        // Mutate Amplitude
        for key in returnIndividual.weightage.keys {
            if Int.random(in: 1...100) > 5 {
                returnIndividual.mutationDirection[key] = 0
                returnIndividual.mutationAmplitude[key] = 1
            }
            if Int.random(in: 1...100) > 50 {
                returnIndividual.mutationDirection[key] = Double.NormalRandom(mu: returnIndividual.mutationDirection[key] ?? 0, sigma: bestError)
                returnIndividual.mutationAmplitude[key]! *= pow(2.71, returnIndividual.mutationDirection[key] ?? 0)
                let mutation = returnIndividual.weightage[key]! * returnIndividual.mutationAmplitude[key]!
                returnIndividual.weightage[key] = mutation
            }
        }
        return returnIndividual
    }

    func Mutate() {
        offspringPopulation = []
        for _ in 1...offspringPopulationSize {
            offspringPopulation.append(Mutation())
        }
    }

    func Selection() {
        let population = (parentPopulation + offspringPopulation + [recombinedEntity]).sorted(by: {
            ($0.error ?? 0) < ($1.error ?? 0)
        })
        offspringPopulation = []
        parentPopulation = Array(population[0..<parentPopulationSize])
    }

    func GenerateAlgorithm(iterationCount : Int) -> ([Double], [Double]) {
        Initialise()
        var bestSolution  = Individual(numberofSignals: 3)
        Evaluate(evaluateParentPopulation: true)
        for iteration in 1...iterationCount {
            Recombine()
            progression.append(recombinedEntity.weightage)
            Mutate()
            Evaluate(evaluateParentPopulation: false)
            Selection()
            if let bstSolution = parentPopulation.sorted(by: {$0.error ?? Double.infinity > $1.error ?? Double.infinity}).first {
                bestSolution = bstSolution
                if bestError > bestSolution.error! {
                    bestError = bestSolution.error!
                }
                errorProgression.append(bestSolution.error!)
                print("Iteration \(iteration): Min Loss \(bestSolution.error!), best fit \(bestSolution.weightage.sorted(by: {$0.key < $1.key}).map({$0.value})).")
            }
            if bestSolution.error ?? 1 == 0 {
                break
            }
        }

        return (data, MixTimeSeries(timeSeries: baseTimeSeries, weightage: bestSolution.weightage))
    }
}