Add parameters (#148)

* first go

* AD

* test

* docstrings

* refactor test

* fix test for GPU

* [skip ci] add basic doc

* add parametric AD tests

* add set_parameter!

* @frapac review changes

* revert copyto! change

* copyto!(@view

* fix doc ref

* no param values in structure

* fix doc ref again

* cleanup registration

* revert type relax

* remove unused

* cleanup registration cont.

* add acopf test

* acopf test gpu-compat

---------

Co-authored-by: Sungho Shin <[email protected]>

Author: klamike

docs/make.jl

@@ -12,6 +12,7 @@ if !(@isdefined _PAGES)
             "guide.md",
             "performance.md",
             "gpu.md",
+            "parameters.md",
             "develop.md",
             "quad.md",
             "distillation.md",
@@ -32,6 +33,7 @@ if !(@isdefined _JL_FILENAMES)
         "opf.jl",
         "gpu.jl",
         "performance.jl",
+        "parameters.jl",
     ]
 end
 

docs/src/parameters.jl

@@ -0,0 +1,44 @@
+# # [Parameters](@id parameters)
+# Parameters act like fixed variables. Internally, ExaModels keeps track of where parameters appear in the model, making it possible to efficiently modify their value without rebuilding the entire model.
+
+# ### Creating Parametric Models
+# 
+# Let's modify the example in [Getting Started](@ref guide) to use parameters. Suppose we want to make the penalty coefficient in the objective function adjustable:
+
+# First, let's create a core:
+using ExaModels, NLPModelsIpopt
+c_param = ExaCore()
+
+# Adding parameters is very similar to adding variables -- just pass a vector of values denoting the initial values.
+θ = parameter(c_param, [100.0, 1.0])  # [penalty_coeff, offset]
+
+# Define the variables as before:
+N = 10
+x_p = variable(c_param, N; start = (mod(i, 2) == 1 ? -1.2 : 1.0 for i = 1:N))
+
+# Now we can use the parameters in our objective function just like variables:
+objective(c_param, θ[1] * (x_p[i-1]^2 - x_p[i])^2 + (x_p[i-1] - θ[2])^2 for i = 2:N)
+
+# Add the same constraints as before:
+constraint(
+    c_param,
+    3x_p[i+1]^3 + 2 * x_p[i+2] - 5 + sin(x_p[i+1] - x_p[i+2])sin(x_p[i+1] + x_p[i+2]) + 4x_p[i+1] -
+    x_p[i]exp(x_p[i] - x_p[i+1]) - 3 for i = 1:(N-2)
+)
+
+# Create the model as before:
+m_param = ExaModel(c_param)
+
+# Solve with original parameters:
+result1 = ipopt(m_param)
+println("Original objective: $(result1.objective)")
+
+# Now change the penalty coefficient and solve again:
+set_parameter!(c_param, θ, [200.0, 1.0])  # Double the penalty coefficient
+result2 = ipopt(m_param)
+println("Modified penalty objective: $(result2.objective)")
+
+# Try a different offset parameter:
+set_parameter!(c_param, θ, [200.0, 0.5])  # Change the offset in the objective
+result3 = ipopt(m_param)
+println("Modified offset objective: $(result3.objective)")