Portfolio-Optimization-Bot/script.jl at main · foxivil/Portfolio-Optimization-Bot · 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
using JuMP
using Gurobi
using CSV
using DataFrames

function portfolio_optimization(path::String, M::Int,  budget::Union{Float64, Int}; analysis::Bool = false, advanced::Bool = false,
    U::Union{Int, Nothing} = nothing, threshold::Union{Float64, Nothing, Int} = nothing, tolerance_level::Union{Float64, Nothing} = nothing, max_violation::Union{Float64, Nothing, Int} = nothing)

    redirect_stdout(devnull)
    df = CSV.read(path, DataFrame)

    projects = df[:, :project]
    profit =  df[:, :profit]
    costs = df[:, :cost]
    risks = df[:, :risk]

    N = length(projects)

    df[!, :dependence] = [
    try
        s === missing || isempty(s) ? nothing : Array(parse.(Int, split(strip(s, ['{', '}']), ",")))
    catch
        nothing
    end
    for s in df.dependence
    ]
    U = isnothing(U) ? N : U
    random_matrix = rand(N, M)
    success_scenarios = random_matrix .> risks;

    model = Model(Gurobi.Optimizer)
    set_optimizer_attribute(model, "OutputFlag", 0)  # Turn off output

    @variable(model, x[1:N], Bin)
    @constraint(model, sum(x[i] for i=1:N) <= U)
    @objective(model, Max, sum(sum(profit[i] * x[i] * success_scenarios[i, j] for j in 1:M) / M for i in 1:N) - sum(costs[i] * x[i] for i in 1:N))
    @constraint(model, sum(costs[i]*x[i] for i in 1:N) <= budget)

    for k in 1:N
        if df[k, :dependence] ≠ nothing
            l = length(df[k, :dependence])
            if l>0
                @constraint(model, sum(x[df[k, :dependence][i]] for i in 1:l) >= l * x[k])
            end
        end
    end

    if advanced
        @variable(model, u[1:M]<=0)
        @variable(model, v[1:M], Bin)
        @constraint(model, [j = 1:M], sum((profit[i]) * x[i] * success_scenarios[i, j] - costs[i]* x[i] for i in 1:N)>= threshold + u[j])
        @constraint(model, [j = 1:M], u[j] >= -v[j] * max_violation)
        @constraint(model, sum( v[j] for j in 1:M)/M <= 1 - tolerance_level)
    end


    optimize!(model)
    status = termination_status(model)

    if status == MOI.OPTIMAL || status == MOI.FEASIBLE_POINT || status == MOI.LOCALLY_SOLVED


        if advanced
            prob = sum(value.(model[:v]))/M
        end

        stockh_x = value.(model[:x]);
        stockh_profit = objective_value(model);
        final_profits = (success_scenarios .* value.(model[:x]))' * profit
        final_costs = sum(value.(model[:x]) .* costs)

        revenue = final_profits .- final_costs

        if analysis
            perfect_information = zeros(N, M)
            for j in 1:M

                model = Model(Gurobi.Optimizer)
                set_optimizer_attribute(model, "OutputFlag", 0)
                @variable(model, x[1:N], Bin)
                @constraint(model, sum(x[i] for i=1:N) <= U)
                @objective(model, Max, sum(profit[i] * x[i] * success_scenarios[i, j] for i in 1:N) - sum(costs[i] * x[i] for i in 1:N))
                @constraint(model, sum(costs[i]*x[i] for i in 1:N) <= budget)
                for k in 1:N
                    if df[k, :dependence] ≠ nothing
                        l = length(df[k, :dependence])
                        if l>0
                            @constraint(model, sum(x[df[k, :dependence][i]] for i in 1:l) >= l * x[k])
                        end
                    end
                end
                optimize!(model)

                perfect_information[:, j] = value.(model[:x])'

            end
            perfect_revenue = (success_scenarios.* perfect_information)' * profit - perfect_information' * costs;
            price_of_perfect_information = sum(perfect_revenue - revenue) * 1/M

            idx = rand(1:M)
            s = success_scenarios[:, idx];

            model = Model(Gurobi.Optimizer)
            set_optimizer_attribute(model, "OutputFlag", 0)
            @variable(model, x[1:N], Bin)
            @constraint(model, sum(x[i] for i=1:N) <= U)
            @objective(model, Max, sum(profit[i] * x[i] * s[i] for i in 1:N) - sum(costs[i] * x[i] for i in 1:N))
            @constraint(model, sum(costs[i]*x[i] for i in 1:N) <= budget)
            for k in 1:N
                if df[k, :dependence] ≠ nothing
                    l = length(df[k, :dependence])
                    if l>0
                        @constraint(model, sum(x[df[k, :dependence][i]] for i in 1:l) >= l * x[k])
                    end
                end
            end
            optimize!(model)
            one_scenario = sum(sum(profit[i] * value.(model[:x])[i] * success_scenarios[i, j] for i in 1:N) for j in 1:M)/M - sum(costs[i] * value.(model[:x])[i] for i in 1:N)
            price_of_scenarios = stockh_profit - one_scenario

            random_matrix_test = rand(N, M)
            success_scenarios_test = random_matrix_test .> risks;

            test_revenue = (success_scenarios_test .* stockh_x)' * profit .- sum(stockh_x .* costs);

            if advanced
                return (stockh_x, stockh_profit, revenue, final_costs, perfect_revenue, price_of_perfect_information,  price_of_scenarios, test_revenue, prob)
            else
                return (stockh_x, stockh_profit, revenue,  final_costs, perfect_revenue, price_of_perfect_information,  price_of_scenarios, test_revenue)
            end
        else
            if advanced
                return (stockh_x, stockh_profit, revenue, final_costs, prob)
            else
                return (stockh_x, stockh_profit, revenue, final_costs)
            end

        end
    else
        println("No feasible solution found.")
        return nothing
    end
end