Tidy up examples and reduce amount of printing in tests (#119)

docs/src/examples.md

@@ -36,7 +36,7 @@ application. Here we categorize the examples by functionality.
 - [Simple Conic](https://github.com/odow/SDDP.jl/tree/master/examples/Conic/simple_conic.jl)
 
 ## Conic
-- [Simple Conic](https://github.com/odow/SDDP.jl/tree/master/examples/Conic/simple_conic.jl)
+- [Simple Conic](https://github.com/odow/SDDP.jl/tree/master/examples/Nonlinear/simple_conic.jl)
 
 ## Termination Criteria
 

examples/Agriculture/mccardle_farm.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -137,7 +137,7 @@ end
 
 srand(111)
 m = mccardle_farm_model()
-solution = solve(m, max_iterations=20)
+solution = solve(m, max_iterations=20, print_level=0)
 @test isapprox(getbound(m), 4074.1391, atol=1e-5)
 
 # results = simulate(m,  # Simulate the policy

examples/AssetManagement/asset_management.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -48,6 +48,8 @@ m = SDDPModel(
 end
 
 srand(111)
-@time solve(m, max_iterations = 25,
-print_level = 0)
+solve(m,
+    max_iterations = 25,
+    print_level    = 0
+)
 @test isapprox(getbound(m), 1.514, atol=1e-4)

examples/AssetManagement/asset_management_stagewise.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -52,7 +52,7 @@ m = SDDPModel(
 end
 
 srand(111)
-@time status = solve(m,
+status = solve(m,
     max_iterations = 100,
     simulation = MonteCarloSimulation(
         frequency = 5,

examples/BinaryProblems/airconditioning.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -51,5 +51,5 @@ end
 
 srand(1234)
 m = airconditioningmodel()
-solve(m, max_iterations=16, print_level=1)
+solve(m, max_iterations=16, print_level=2)
 @test isapprox(getbound(m), 62_500.0)

examples/BinaryProblems/booking_management.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Eyob Zewdie, Oscar Dowson
+#  Copyright 2018, Eyob Zewdie, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -88,5 +88,5 @@ m_1_2_5 = bookingmanagementmodel(1, 2, 5)
 
 srand(1234)
 m_2_2_3 = bookingmanagementmodel(2, 2, 3)
-@test solve(m_2_2_3, max_iterations = 40, print_level = 2) == :max_iterations
+@test solve(m_2_2_3, max_iterations = 40, print_level = 0) == :max_iterations
 @test isapprox(getbound(m_2_2_3), 6.13, atol=0.001)

examples/BinaryProblems/vehicle_location.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017,  Oscar Dowson, Eyob Zewdie
+#  Copyright 2018,  Oscar Dowson, Eyob Zewdie
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -93,7 +93,7 @@ function vehiclelocationmodel(nvehicles, baselocations, requestlocations)
 end
 
 ambulancemodel = vehiclelocationmodel(3, [0, 20, 40, 60, 80, 100], 0:2:100)
-@test solve(ambulancemodel, max_iterations=50) == :max_iterations
+@test solve(ambulancemodel, max_iterations=50, print_level=0) == :max_iterations
 @test isapprox(getbound(ambulancemodel), 1700.0, atol=5)
 # N = 100
 # results = simulate(m, N, [:dispatch, :q, :shift])

examples/HydroValleys/hydro_valley_tests.jl

@@ -13,7 +13,7 @@ include(joinpath(dirname(@__FILE__), "hydro_valley.jl"))
 
 # deterministic
 deterministic_model = hydrovalleymodel(hasmarkovprice=false,hasstagewiseinflows=false)
-SDDP.solve(deterministic_model, max_iterations=10, cut_selection_frequency=1)
+SDDP.solve(deterministic_model, max_iterations=10, cut_selection_frequency=1, print_level=0)
 @test isapprox(getbound(deterministic_model), 835.0, atol=1e-3)
 
 # stagewise inflows
@@ -50,7 +50,8 @@ SDDP.solve(worst_case_model,
         min  = 20,
         step = 10,
         max  = 50
-    )
+    ),
+    print_level=0
 )
 
 @test isapprox(getbound(worst_case_model), -780.867, atol=1e-3)
@@ -68,10 +69,10 @@ SDDP.solve(cutselection_model,
 
 # Distributionally robust Optimization
 dro_model = hydrovalleymodel(hasmarkovprice=false, riskmeasure=DRO(sqrt(2/3)-1e-6))
-SDDP.solve(dro_model, max_iterations = 10)
+SDDP.solve(dro_model, max_iterations = 10, print_level=0)
 @test isapprox(getbound(dro_model), 835.0, atol=1e-3)
 dro_model = hydrovalleymodel(hasmarkovprice=false, riskmeasure=DRO(1/6))
-SDDP.solve(dro_model, max_iterations = 20)
+SDDP.solve(dro_model, max_iterations = 20, print_level=0)
 @test isapprox(getbound(dro_model), 836.695, atol=1e-3)
 # (Note) radius ≈ sqrt(2/3), will set all noise probabilities to zero except the worst case noise
 # (Why?):

examples/Newsvendor/async_newsvendor.jl

@@ -86,7 +86,7 @@ end
 m = createmodel(EAVaR(beta   = 0.6, lambda = 0.5))
 
 
-@test_throws Exception SDDP.solve(m, max_iterations=30, solve_type=Asynchronous(slaves=[2]))
+@test_throws Exception SDDP.solve(m, print_level=0, max_iterations=30, solve_type=Asynchronous(slaves=[2]))
 
 # slave processes
 slaves = workers()

examples/Newsvendor/newsvendor.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -95,7 +95,7 @@ news1 = newsvendormodel()
 news2 = newsvendormodel(riskmeasure=EAVaR(beta=0.6,lambda=0.5))
 @test SDDP.solve(news2,
     max_iterations = 50,
-    print_level = 0,
+    print_level    = 0,
     simulation     = MonteCarloSimulation(
                         frequency = 10,
                         min       = 500,

examples/Conic/inventory-control.jl → examples/Nonlinear/inventory-control.jl

@@ -10,12 +10,12 @@
         Volume I (3rd ed.). Bellmont, MA: Athena Scientific.
 =#
 
-using SDDP, JuMP, Gurobi, Base.Test
+using SDDP, JuMP, Ipopt, Base.Test
 
 m = SDDPModel(
         stages          = 3,
         sense           = :Min,
-        solver          = GurobiSolver(OutputFlag=0),
+        solver          = IpoptSolver(print_level=0),
         objective_bound = 0.0
             ) do sp, t
     @state(sp, 0 <= x <= 2, x0 == 0)

examples/Nonlinear/logs.jl

@@ -0,0 +1,27 @@
+#  Copyright 2018, Oscar Dowson
+#  This Source Code Form is subject to the terms of the Mozilla Public
+#  License, v. 2.0. If a copy of the MPL was not distributed with this
+#  file, You can obtain one at http://mozilla.org/MPL/2.0/.
+#############################################################################
+
+using JuMP, SDDP, Ipopt
+
+m = SDDPModel(
+    stages          = 2,
+    sense           = :Max,
+    objective_bound = 2.0,
+    solver          = IpoptSolver(print_level=0)
+        ) do sp, t
+
+    @state(sp, x′ >= 1, x==1)
+    if t == 1
+        @rhsnoise(sp, w=[1,2,3], x′ == w)
+        @stageobjective(sp, 0.0)
+    else
+        @variable(sp, y)
+        @NLconstraint(sp, y <= log(x))
+        @stageobjective(sp, y)
+    end
+end
+solve(m, max_iterations=10, print_level=0)
+@test isapprox(getbound(m), log(6) / 3, atol=1e-4)

examples/Conic/simple_conic.jl → examples/Nonlinear/simple_conic.jl

@@ -1,10 +1,10 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
 #############################################################################
 
-using SDDP, JuMP, Gurobi, Base.Test
+using SDDP, JuMP, Ipopt, Base.Test
 
 """
 A trivial two-stage stochastic programming problem to work out the
@@ -13,11 +13,11 @@ and the second-stage evaluates the decision.
 """
 function least_squares(data::Vector{Float64},radius=0.0)
     m = SDDPModel(
-        stages = 2,
-        sense = :Min,
+        stages          = 2,
+        sense           = :Min,
         objective_bound = 0.0,
-        solver=GurobiSolver(OutputFlag=0),
-        risk_measure = DRO(radius)
+        solver          = IpoptSolver(print_level=0),
+        risk_measure    = DRO(radius)
             ) do sp, t
 
         @state(sp, x′>=0, x==0)

examples/PriceInterpolation/simple_contracting.jl

@@ -125,7 +125,7 @@ end
 # dynamic interpolation
 m = contracting_example()
 srand(123)
-SDDP.solve(m, max_iterations = 50, cut_selection_frequency=10, print_level=2)
+SDDP.solve(m, max_iterations = 50, cut_selection_frequency=10, print_level=0)
 @test SDDP.getbound(m) <= 175.0
 
 # historical simulation
@@ -153,7 +153,7 @@ results = simulate(m3, 500)
 # 5 fixed ribs
 m5 = contracting_example(5)
 srand(123)
-SDDP.solve(m5, max_iterations = 10)
+SDDP.solve(m5, max_iterations = 10, print_level=0)
 @test SDDP.getbound(m5) <= 150.0
 
 # historical simulation

examples/PriceInterpolation/widget_producer.jl

@@ -145,15 +145,16 @@ end
 m = widget_producer_example()
 srand(123)
 status = SDDP.solve(m,
-    time_limit     = 20.0,
+    time_limit     = 5.0,
     simulation = MonteCarloSimulation(
         frequency = 100, min=100, step=100,max=1000
-    )
+    ),
+    print_level = 0
 )
 @test status == :time_limit
 results = simulate(m, 1000, [:C, :S, :price, :s, :d, :b])
 @test length(results) == 1000
-@test getbound(m) <= 12.0
+@test getbound(m) <= 15.0
 
 # SDDP.save!("results.julia", results)
 # plt = SDDP.newplot()

examples/StochDynamicProgramming.jl/multistock-example.jl

@@ -1,5 +1,5 @@
 #==
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -46,7 +46,7 @@ m = SDDPModel(
     @stageobjective(sp, (sin(3 * stage) - 1) * sum(control))
 end
 
-@time status = SDDP.solve(m, max_iterations = 100,
+status = SDDP.solve(m, max_iterations = 100,
 print_level = 0)
 @test isapprox(SDDP.getbound(m), -4.349, atol=0.01)
 

examples/StochDynamicProgramming.jl/stock-example.jl

@@ -1,5 +1,5 @@
 #==
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -39,7 +39,7 @@ m = SDDPModel(
     @stageobjective(sp, (sin(3 * stage) - 1) * control)
 end
 
-@time status = SDDP.solve(m, max_iterations = 50,
+status = SDDP.solve(m, max_iterations = 50,
 print_level = 0)
 @test isapprox(SDDP.getbound(m), -1.471, atol=0.001)
 

examples/StructDualDynProg.jl/prob5.2_2stages.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -10,22 +10,22 @@
 
 using SDDP, JuMP, Base.Test, Clp
 
-n = 4
-m = 3
-ic = [16, 5, 32, 2]
-C = [25, 80, 6.5, 160]
-T = [8760, 7000, 1500] / 8760
-D2 = [diff([0, 3919, 7329, 10315])  diff([0, 7086, 9004, 11169])]
-p2 = [0.9, 0.1]
-
 mod = SDDPModel(
                   sense = :Min,
                  stages = 2,
                  solver = ClpSolver(),
-        objective_bound = 0,
-        # noise_probability = [ Float64[], p2, p2 ]
+        objective_bound = 0
                                 ) do sp, t
 
+    n = 4
+    m = 3
+    ic = [16, 5, 32, 2]
+    C = [25, 80, 6.5, 160]
+    T = [8760, 7000, 1500] / 8760
+    D2 = [diff([0, 3919, 7329, 10315])  diff([0, 7086, 9004, 11169])]
+    p2 = [0.9, 0.1]
+
+
     @state(sp, x[i=1:n] >= 0, x0 == 0)
 
     @variables(sp, begin
@@ -52,10 +52,10 @@ mod = SDDPModel(
     end
 end
 
-@time status = SDDP.solve(mod,
+status = SDDP.solve(mod,
     max_iterations = 50,
-    print_level = 0,
-    simulation = MonteCarloSimulation(
+    print_level    = 0,
+    simulation     = MonteCarloSimulation(
         frequency = 10,
         min       = 100,
         step      = 1,

examples/StructDualDynProg.jl/prob5.2_3stages.jl

@@ -1,4 +1,4 @@
-#  Copyright 2017, Oscar Dowson
+#  Copyright 2018, Oscar Dowson
 #  This Source Code Form is subject to the terms of the Mozilla Public
 #  License, v. 2.0. If a copy of the MPL was not distributed with this
 #  file, You can obtain one at http://mozilla.org/MPL/2.0/.
@@ -10,20 +10,19 @@
 
 using SDDP, JuMP, Base.Test, Clp
 
-n = 4
-m = 3
-ic = [16, 5, 32, 2]
-C = [25, 80, 6.5, 160]
-T = [8760, 7000, 1500] / 8760
-D2 = [diff([0, 3919, 7329, 10315])  diff([0, 7086, 9004, 11169])]
-p2 = [0.9, 0.1]
-
 mod = SDDPModel(
                   sense = :Min,
                  stages = 3,
                  solver = ClpSolver(),
         objective_bound = 0
                                 ) do sp, t
+    n = 4
+    m = 3
+    ic = [16, 5, 32, 2]
+    C = [25, 80, 6.5, 160]
+    T = [8760, 7000, 1500] / 8760
+    D2 = [diff([0, 3919, 7329, 10315])  diff([0, 7086, 9004, 11169])]
+    p2 = [0.9, 0.1]
 
     @state(sp, x[i=1:n] >= 0, x0 == 0)
 
@@ -52,10 +51,10 @@ mod = SDDPModel(
     end
 end
 
-@time status = SDDP.solve(mod,
+status = SDDP.solve(mod,
     max_iterations = 30,
-    print_level = 0,
-    simulation = MonteCarloSimulation(
+    print_level    = 0,
+    simulation     = MonteCarloSimulation(
         frequency = 10,
         min       = 100,
         step      = 1,