Merge pull request #2208 from SciML/myb/sdde

SDDE support

Project.toml

@@ -114,9 +114,10 @@ SafeTestsets = "1bc83da4-3b8d-516f-aca4-4fe02f6d838f"
 StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 SteadyStateDiffEq = "9672c7b4-1e72-59bd-8a11-6ac3964bc41f"
+StochasticDelayDiffEq = "29a0d76e-afc8-11e9-03a4-eda52ae4b960"
 StochasticDiffEq = "789caeaf-c7a9-5a7d-9973-96adeb23e2a0"
 Sundials = "c3572dad-4567-51f8-b174-8c6c989267f4"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["AmplNLWriter", "BenchmarkTools", "ControlSystemsMTK", "NonlinearSolve", "ForwardDiff", "Ipopt", "Ipopt_jll", "ModelingToolkitStandardLibrary", "Optimization", "OptimizationOptimJL", "OptimizationMOI", "OrdinaryDiffEq", "Random", "ReferenceTests", "SafeTestsets", "StableRNGs", "Statistics", "SteadyStateDiffEq", "Test", "StochasticDiffEq", "Sundials"]
+test = ["AmplNLWriter", "BenchmarkTools", "ControlSystemsMTK", "NonlinearSolve", "ForwardDiff", "Ipopt", "Ipopt_jll", "ModelingToolkitStandardLibrary", "Optimization", "OptimizationOptimJL", "OptimizationMOI", "OrdinaryDiffEq", "Random", "ReferenceTests", "SafeTestsets", "StableRNGs", "Statistics", "SteadyStateDiffEq", "Test", "StochasticDiffEq", "Sundials", "StochasticDelayDiffEq"]

src/systems/diffeqs/abstractodesystem.jl

@@ -169,14 +169,14 @@ function isdelay(var, iv)
     return false
 end
 const DDE_HISTORY_FUN = Sym{Symbolics.FnType{Tuple{Any, <:Real}, Vector{Real}}}(:___history___)
-function delay_to_function(sys::AbstractODESystem)
-    delay_to_function(full_equations(sys),
+function delay_to_function(sys::AbstractODESystem, eqs = full_equations(sys))
+    delay_to_function(eqs,
         get_iv(sys),
         Dict{Any, Int}(operation(s) => i for (i, s) in enumerate(states(sys))),
         parameters(sys),
         DDE_HISTORY_FUN)
 end
-function delay_to_function(eqs::Vector{<:Equation}, iv, sts, ps, h)
+function delay_to_function(eqs::Vector, iv, sts, ps, h)
     delay_to_function.(eqs, (iv,), (sts,), (ps,), (h,))
 end
 function delay_to_function(eq::Equation, iv, sts, ps, h)
@@ -548,8 +548,8 @@ function DiffEqBase.DDEFunction{iip}(sys::AbstractODESystem, dvs = states(sys),
         expression_module = eval_module, checkbounds = checkbounds,
         kwargs...)
     f_oop, f_iip = (drop_expr(@RuntimeGeneratedFunction(eval_module, ex)) for ex in f_gen)
-    f(u, p, h, t) = f_oop(u, p, h, t)
-    f(du, u, p, h, t) = f_iip(du, u, p, h, t)
+    f(u, h, p, t) = f_oop(u, h, p, t)
+    f(du, u, h, p, t) = f_iip(du, u, h, p, t)
 
     DDEFunction{iip}(f,
         sys = sys,
@@ -558,6 +558,35 @@ function DiffEqBase.DDEFunction{iip}(sys::AbstractODESystem, dvs = states(sys),
         paramsyms = Symbol.(ps))
 end
 
+function DiffEqBase.SDDEFunction(sys::AbstractODESystem, args...; kwargs...)
+    SDDEFunction{true}(sys, args...; kwargs...)
+end
+
+function DiffEqBase.SDDEFunction{iip}(sys::AbstractODESystem, dvs = states(sys),
+    ps = parameters(sys), u0 = nothing;
+    eval_module = @__MODULE__,
+    checkbounds = false,
+    kwargs...) where {iip}
+    f_gen = generate_function(sys, dvs, ps; isdde = true,
+        expression = Val{true},
+        expression_module = eval_module, checkbounds = checkbounds,
+        kwargs...)
+    f_oop, f_iip = (drop_expr(@RuntimeGeneratedFunction(eval_module, ex)) for ex in f_gen)
+    g_gen = generate_diffusion_function(sys, dvs, ps; expression = Val{true},
+        isdde = true, kwargs...)
+    g_oop, g_iip = (drop_expr(@RuntimeGeneratedFunction(ex)) for ex in g_gen)
+    f(u, h, p, t) = f_oop(u, h, p, t)
+    f(du, u, h, p, t) = f_iip(du, u, h, p, t)
+    g(u, h, p, t) = g_oop(u, h, p, t)
+    g(du, u, h, p, t) = g_iip(du, u, h, p, t)
+
+    SDDEFunction{iip}(f, g,
+        sys = sys,
+        syms = Symbol.(dvs),
+        indepsym = Symbol(get_iv(sys)),
+        paramsyms = Symbol.(ps))
+end
+
 """
 ```julia
 ODEFunctionExpr{iip}(sys::AbstractODESystem, dvs = states(sys),
@@ -941,6 +970,72 @@ function DiffEqBase.DDEProblem{iip}(sys::AbstractODESystem, u0map = [],
     DDEProblem{iip}(f, u0, h, tspan, p; kwargs1..., kwargs...)
 end
 
+function DiffEqBase.SDDEProblem(sys::AbstractODESystem, args...; kwargs...)
+    SDDEProblem{true}(sys, args...; kwargs...)
+end
+function DiffEqBase.SDDEProblem{iip}(sys::AbstractODESystem, u0map = [],
+    tspan = get_tspan(sys),
+    parammap = DiffEqBase.NullParameters();
+    callback = nothing,
+    check_length = true,
+    sparsenoise = nothing,
+    kwargs...) where {iip}
+    has_difference = any(isdifferenceeq, equations(sys))
+    f, u0, p = process_DEProblem(SDDEFunction{iip}, sys, u0map, parammap;
+        t = tspan !== nothing ? tspan[1] : tspan,
+        has_difference = has_difference,
+        symbolic_u0 = true,
+        check_length, kwargs...)
+    h_oop, h_iip = generate_history(sys, u0)
+    h(out, p, t) = h_iip(out, p, t)
+    h(p, t) = h_oop(p, t)
+    u0 = h(p, tspan[1])
+    cbs = process_events(sys; callback, has_difference, kwargs...)
+    if has_discrete_subsystems(sys) && (dss = get_discrete_subsystems(sys)) !== nothing
+        affects, clocks, svs = ModelingToolkit.generate_discrete_affect(dss...)
+        discrete_cbs = map(affects, clocks, svs) do affect, clock, sv
+            if clock isa Clock
+                PeriodicCallback(DiscreteSaveAffect(affect, sv), clock.dt)
+            else
+                error("$clock is not a supported clock type.")
+            end
+        end
+        if cbs === nothing
+            if length(discrete_cbs) == 1
+                cbs = only(discrete_cbs)
+            else
+                cbs = CallbackSet(discrete_cbs...)
+            end
+        else
+            cbs = CallbackSet(cbs, discrete_cbs)
+        end
+    else
+        svs = nothing
+    end
+    kwargs = filter_kwargs(kwargs)
+
+    kwargs1 = (;)
+    if cbs !== nothing
+        kwargs1 = merge(kwargs1, (callback = cbs,))
+    end
+    if svs !== nothing
+        kwargs1 = merge(kwargs1, (disc_saved_values = svs,))
+    end
+
+    noiseeqs = get_noiseeqs(sys)
+    sparsenoise === nothing && (sparsenoise = get(kwargs, :sparse, false))
+    if noiseeqs isa AbstractVector
+        noise_rate_prototype = nothing
+    elseif sparsenoise
+        I, J, V = findnz(SparseArrays.sparse(noiseeqs))
+        noise_rate_prototype = SparseArrays.sparse(I, J, zero(eltype(u0)))
+    else
+        noise_rate_prototype = zeros(eltype(u0), size(noiseeqs))
+    end
+    SDDEProblem{iip}(f, f.g, u0, h, tspan, p; noise_rate_prototype =
+        noise_rate_prototype, kwargs1..., kwargs...)
+end
+
 """
 ```julia
 ODEProblemExpr{iip}(sys::AbstractODESystem, u0map, tspan,

src/systems/diffeqs/sdesystem.jl

@@ -210,11 +210,18 @@ function Base.:(==)(sys1::SDESystem, sys2::SDESystem)
 end
 
 function generate_diffusion_function(sys::SDESystem, dvs = states(sys),
-    ps = parameters(sys); kwargs...)
-    return build_function(get_noiseeqs(sys),
-        map(x -> time_varying_as_func(value(x), sys), dvs),
-        map(x -> time_varying_as_func(value(x), sys), ps),
-        get_iv(sys); kwargs...)
+    ps = parameters(sys); isdde = false, kwargs...)
+    eqs = get_noiseeqs(sys)
+    if isdde
+        eqs = delay_to_function(sys, eqs)
+    end
+    u = map(x -> time_varying_as_func(value(x), sys), dvs)
+    p = map(x -> time_varying_as_func(value(x), sys), ps)
+    if isdde
+        return build_function(eqs, u, DDE_HISTORY_FUN, p, get_iv(sys); kwargs...)
+    else
+        return build_function(eqs, u, p, get_iv(sys); kwargs...)
+    end
 end
 
 """

test/dde.jl

@@ -49,3 +49,36 @@ prob2 = DDEProblem(sys,
     constant_lags = [tau])
 sol2_mtk = solve(prob2, alg, reltol = 1e-7, abstol = 1e-10)
 @test sol2_mtk.u[end] ≈ sol2.u[end]
+
+using StochasticDelayDiffEq
+function hayes_modelf(du, u, h, p, t)
+    τ, a, b, c, α, β, γ = p
+    du .= a .* u .+ b .* h(p, t - τ) .+ c
+end
+function hayes_modelg(du, u, h, p, t)
+    τ, a, b, c, α, β, γ = p
+    du .= α .* u .+ γ
+end
+h(p, t) = (ones(1) .+ t);
+tspan = (0.0, 10.0)
+
+pmul = [1.0,
+    -4.0, -2.0, 10.0,
+    -1.3, -1.2, 1.1]
+
+prob = SDDEProblem(hayes_modelf, hayes_modelg, [1.0], h, tspan, pmul;
+    constant_lags = (pmul[1],));
+sol = solve(prob, RKMil())
+
+@variables t x(..)
+@parameters a=-4.0 b=-2.0 c=10.0 α=-1.3 β=-1.2 γ=1.1
+D = Differential(t)
+@brownian η
+τ = 1.0
+eqs = [D(x(t)) ~ a * x(t) + b * x(t - τ) + c + (α * x(t) + γ) * η]
+@named sys = System(eqs)
+sys = structural_simplify(sys)
+@test equations(sys) == [D(x(t)) ~ a * x(t) + b * x(t - τ) + c]
+@test isequal(ModelingToolkit.get_noiseeqs(sys), [α * x(t) + γ;;])
+prob_mtk = SDDEProblem(sys, [x(t) => 1.0 + t], tspan; constant_lags = (τ,));
+@test_nowarn sol_mtk = solve(prob_mtk, RKMil())