Merge pull request #203 from lanl-ansi/Alpine.jl-Link

New Feature: Linking Constraints for Multilinear Terms

.JuliaFormatter.toml

@@ -0,0 +1,8 @@
+# Configuration file for JuliaFormatter.jl
+ # For more information, see: https://domluna.github.io/JuliaFormatter.jl/stable/config/
+
+ always_for_in = true
+ always_use_return = true
+ margin = 90
+ remove_extra_newlines = true
+ short_to_long_function_def = true

.github/workflows/format_check.yml

@@ -0,0 +1,32 @@
+name: format-check
+on:
+  push:
+    branches:
+      - master
+      - release-*
+  pull_request:
+    types: [opened, synchronize, reopened]
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: julia-actions/setup-julia@latest
+        with:
+          version: '1'
+      - uses: actions/checkout@v1
+      - name: Format check
+        shell: julia --color=yes {0}
+        run: |
+          using Pkg
+          Pkg.add(PackageSpec(name="JuliaFormatter", version="1"))
+          using JuliaFormatter
+          format("src/MOI_wrapper", verbose=true)
+          format("test", verbose=true)
+          format("docs", verbose=true)
+          out = String(read(Cmd(`git diff`)))
+          if isempty(out)
+              exit(0)
+          end
+          @error "Some files have not been formatted !!!"
+          write(stdout, out)
+          exit(1)

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Alpine.jl Change Log
 
+## v0.5.0
+- New feature: Linking constraints for multilinear terms with uniform and adaptive partitions (significant speed up in run times for multilinear problems: http://www.optimization-online.org/DB_HTML/2022/07/8974.html) (@jongeunkim)
+- Added dependency on Combinatorics package for linking constraints 
+- Added unit tests for linking constraints feature
+- Changed Cbc test dependency to HiGHS solver
+- Dropped a few redundant unit tests
+- Dropped support for (slower) piecewise McCormick formulations
+- Changed `disc_ratio` to `partition_scaling_factor`
+- Added JuliaFormatter.toml and formatting workfow
+
 ## v0.4.3
 - Dropped support for checking solver availability. User can choose to use any MILP, MINLP and/or NLP solver with Alpine
 - `GRB_ENV` now works with Gurobi solver to avoid printing License info multiple times during iterations
@@ -185,7 +195,7 @@ Closed issues:
 - Alpine.jl first release
 
 ## Merged Features
-- `disc_ratio` selection algorithm
+- `partition_scaling_factor` selection algorithm
 - logarithm embedding formulation
 - bounding model warm starting and no-good-cuts using solution pool
 - `disc_vars` selection algorithm based on weighted minimum vertex cover
@@ -229,6 +239,6 @@ Closed issues:
 - Fixed many other things
 
 ## Tests & Coverage
-- Most new features are covered (embedding, `disc_ratio` algorithm, etc.)
+- Most new features are covered (embedding, `partition_scaling_factor` algorithm, etc.)
 - Stronger expression parsing tests with new operators covered
 - Coverage should be more than `~90%`

Project.toml

@@ -2,30 +2,32 @@ name = "Alpine"
 uuid = "07493b3f-dabb-5b16-a503-4139292d7dd4"
 authors = ["Harsha Nagarajan, Site Wang, Kaarthik Sundar and contributors"]
 repo = "https://github.com/lanl-ansi/Alpine.jl.git"
-version = "0.4.3"
+version = "0.5.0"
 
 [deps]
+Combinatorics = "861a8166-3701-5b0c-9a16-15d98fcdc6aa"
 JuMP = "4076af6c-e467-56ae-b986-b466b2749572"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 
 [compat]
-Cbc = "0.8, 0.9, 1"
+HiGHS = "1"
 Ipopt = "0.8, 0.9, 1"
 JuMP = "0.22, 0.23, 1"
 Juniper = "0.8, 0.9"
 MathOptInterface = "0.10, 1"
 Pavito = "0.3.4, 0.3.5"
 julia = "1"
+Combinatorics = "1"
 
 [extras]
-Cbc = "9961bab8-2fa3-5c5a-9d89-47fab24efd76"
+HiGHS = "87dc4568-4c63-4d18-b0c0-bb2238e4078b"
 Ipopt = "b6b21f68-93f8-5de0-b562-5493be1d77c9"
 Juniper = "2ddba703-00a4-53a7-87a5-e8b9971dde84"
 Pavito = "cd433a01-47d1-575d-afb7-6db927ee8d8f"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Cbc", "Ipopt", "Juniper", "Pavito", "Random", "Test"]
+test = ["HiGHS", "Ipopt", "Juniper", "Pavito", "Random", "Test"]

README.md

@@ -63,8 +63,8 @@ We are seeking out hard benchmark instances for MINLPs. Please get in touch eith
 If you find Alpine useful in your work, we kindly request that you cite the following papers ([pdf](http://harshangrjn.github.io/pdf/JOGO_2018.pdf), [pdf](http://harshangrjn.github.io/pdf/CP_2016.pdf))
 ```bibtex
 @article{alpine_JOGO2019,
-  author = {Nagarajan, Harsha and Lu, Mowen and Wang, Site and Bent, Russell and Sundar, Kaarthik},
   title = {An adaptive, multivariate partitioning algorithm for global optimization of nonconvex programs},
+  author = {Nagarajan, Harsha and Lu, Mowen and Wang, Site and Bent, Russell and Sundar, Kaarthik},
   journal = {Journal of Global Optimization},
   year = {2019},
   issn = {1573-2916},
@@ -81,7 +81,7 @@ If you find Alpine useful in your work, we kindly request that you cite the foll
   doi = {10.1007/978-3-319-44953-1_24},
 }
 ```
-If you find the underlying piecewise polyhedral formulations implemented in Alpine useful in your work, we kindly request that you cite the following paper ([link](https://doi.org/10.1016/j.orl.2020.12.002)): 
+If you find the underlying piecewise polyhedral formulations implemented in Alpine useful in your work, we kindly request that you cite the following papers ([link-1](https://doi.org/10.1016/j.orl.2020.12.002), [link-2](http://www.optimization-online.org/DB_HTML/2022/07/8974.html)): 
 ```bibtex
 @article{alpine_ORL2021,
   title = {Piecewise polyhedral formulations for a multilinear term},
@@ -93,4 +93,11 @@ If you find the underlying piecewise polyhedral formulations implemented in Alpi
   year = {2021},
   publisher = {Elsevier}
 }
+
+@article{alpine_OptOnline2022,
+    title={Piecewise Polyhedral Relaxations of Multilinear Optimization},
+    author={Kim, Jongeun and Richard, Jean-Philippe P. and Tawarmalani, Mohit},
+    eprinttype={Optimization Online},
+    date={2022}
+}
 ```

docs/make.jl

@@ -5,17 +5,25 @@ using DocumenterTools: Themes
 for w in ("light", "dark")
     header = read(joinpath(@__DIR__, "src/assets/themes/style.scss"), String)
     theme = read(joinpath(@__DIR__, "src/assets/themes/$(w)defs.scss"), String)
-    write(joinpath(@__DIR__, "src/assets/themes/$(w).scss"), header*"\n"*theme)
+    write(joinpath(@__DIR__, "src/assets/themes/$(w).scss"), header * "\n" * theme)
 end
 
-Themes.compile(joinpath(@__DIR__, "src/assets/themes/light.scss"), joinpath(@__DIR__, "src/assets/themes/documenter-light.css"))
-Themes.compile(joinpath(@__DIR__, "src/assets/themes/dark.scss"), joinpath(@__DIR__, "src/assets/themes/documenter-dark.css"))
+Themes.compile(
+    joinpath(@__DIR__, "src/assets/themes/light.scss"),
+    joinpath(@__DIR__, "src/assets/themes/documenter-light.css"),
+)
+Themes.compile(
+    joinpath(@__DIR__, "src/assets/themes/dark.scss"),
+    joinpath(@__DIR__, "src/assets/themes/documenter-dark.css"),
+)
 
 makedocs(
     sitename = "Alpine",
     authors = "Harsha Nagarajan, Site Wang, Kaarthik Sundar and contributors",
-    format = Documenter.HTML(mathengine = Documenter.MathJax(), 
-                             prettyurls = get(ENV, "CI", nothing) == "true"),
+    format = Documenter.HTML(
+        mathengine = Documenter.MathJax(),
+        prettyurls = get(ENV, "CI", nothing) == "true",
+    ),
     strict = true,
     pages = [
         "Introduction" => "index.md",
@@ -26,8 +34,4 @@ makedocs(
     ],
 )
 
-deploydocs(
-    repo = "github.com/lanl-ansi/Alpine.jl.git",
-    push_preview = true
-)
-
+deploydocs(repo = "github.com/lanl-ansi/Alpine.jl.git", push_preview = true)

docs/src/index.md

@@ -20,7 +20,7 @@ import Pkg
 Pkg.add("Alpine")
 ```
 
-At least one local nonlinear-programming (NLP) solver and a convex mixed-integer programming (MIP) solver are required for running Alpine. If your problem is an NLP, we recommend [Ipopt](https://github.com/jump-dev/Ipopt.jl), and else if your problem is a mixed-integer NLP (MINLP), we recommend [Juniper](https://github.com/lanl-ansi/Juniper.jl). For the underlying MIP solver, [Gurobi](https://github.com/jump-dev/Gurobi.jl) is highly recommended, as it is fast, scaleable and can be used to solve on fairly large-scale non-convex programs. However, the open-source [CBC](https://github.com/jump-dev/Cbc.jl) or [GLPK](https://github.com/jump-dev/GLPK.jl) solver is also compatible with Alpine. For example, Ipopt and GLPK can be installed via the package manager with
+At least one local nonlinear-programming (NLP) solver and a convex mixed-integer programming (MIP) solver are required for running Alpine. If your problem is an NLP, we recommend [Ipopt](https://github.com/jump-dev/Ipopt.jl), and else if your problem is a mixed-integer NLP (MINLP), we recommend [Juniper](https://github.com/lanl-ansi/Juniper.jl). For the underlying MIP solver, [Gurobi](https://github.com/jump-dev/Gurobi.jl) is highly recommended, as it is fast, scaleable and can be used to solve on fairly large-scale non-convex programs. However, the open-source [HiGHS](https://github.com/jump-dev/HiGHS.jl) or [GLPK](https://github.com/jump-dev/GLPK.jl) solver is also compatible with Alpine. For example, Ipopt and Gurobi can be installed via the package manager with
 
 ```julia
 import Pkg
@@ -38,14 +38,44 @@ Pkg.test("Alpine")
 ```
 
 ## Citing Alpine
-If you find Alpine useful in your work, we request you to cite the following paper [\[link\]](https://harshangrjn.github.io/pdf/JOGO_2018.pdf): 
+If you find Alpine useful in your work, we request you to cite the following paper [\[pdf\]](https://harshangrjn.github.io/pdf/JOGO_2018.pdf) [\[pdf\]](http://harshangrjn.github.io/pdf/CP_2016.pdf): 
 ```bibtex
 @article{alpine_JOGO2019,
-  author = {Nagarajan, Harsha and Lu, Mowen and Wang, Site and Bent, Russell and Sundar, Kaarthik},
   title = {An adaptive, multivariate partitioning algorithm for global optimization of nonconvex programs},
+  author = {Nagarajan, Harsha and Lu, Mowen and Wang, Site and Bent, Russell and Sundar, Kaarthik},
   journal = {Journal of Global Optimization},
   year = {2019},
   issn = {1573-2916},
   doi = {10.1007/s10898-018-00734-1},
 }
+
+@inproceedings{alpine_CP2016,
+  title = {Tightening {McCormick} relaxations for nonlinear programs via dynamic multivariate partitioning},
+  author = {Nagarajan, Harsha and Lu, Mowen and Yamangil, Emre and Bent, Russell},
+  booktitle = {International Conference on Principles and Practice of Constraint Programming},
+  pages = {369--387},
+  year = {2016},
+  organization = {Springer},
+  doi = {10.1007/978-3-319-44953-1_24},
+}
+```
+If you find the underlying piecewise polyhedral formulations implemented in Alpine useful in your work, we kindly request that you cite the following papers ([link-1](https://doi.org/10.1016/j.orl.2020.12.002), [link-2](http://www.optimization-online.org/DB_HTML/2022/07/8974.html)): 
+```bibtex
+@article{alpine_ORL2021,
+  title = {Piecewise polyhedral formulations for a multilinear term},
+  author = {Sundar, Kaarthik and Nagarajan, Harsha and Linderoth, Jeff and Wang, Site and Bent, Russell},
+  journal = {Operations Research Letters},
+  volume = {49},
+  number = {1},
+  pages = {144--149},
+  year = {2021},
+  publisher = {Elsevier}
+}
+
+@article{alpine_OptOnline2022,
+    title={Piecewise Polyhedral Relaxations of Multilinear Optimization},
+    author={Kim, Jongeun and Richard, Jean-Philippe P. and Tawarmalani, Mohit},
+    eprinttype={Optimization Online},
+    date={2022}
+}
 ```

docs/src/parameters.md

@@ -21,7 +21,7 @@ Here are a few general solver options which control the performance of Alpine:
 ## Adaptive Partitioning Options
 * `apply_partitioning (default = true)`: applies Alpine's built-in MIP-based partitioning algorithm only when activated; else terminates with the presolve solution.
 
-* `disc_ratio (default = 4)`: used during [`add_adaptive_partition`](@ref) for measuring the width of new partitions relative to the active partition chosen in the sequentially solved lower-bounding MIPs. This value can substantially affect the run time for global convergence; this value can be set to different integer values (>= 4) for various classes of problems. 
+* `partition_scaling_factor (default = 10)`: used during [`add_adaptive_partition`](@ref) for scaling the width of new partitions relative to the active partition chosen in the sequentially-solved lower-bounding MIP models. This value can substantially affect the run time for global convergence; this value can be set to different integer values (>= 4) for various classes of problems. 
 
 * `disc_var_pick (default = 0)`: controls Alpine's algorithm used for selecting variables for partitioning; `0` is for max-cover, `1` is for minimum-vertex-cover. This parameter allows functional inputs.
 

examples/JuMP_models.jl

@@ -9,6 +9,5 @@ include("MINLPs/div.jl")
 include("MINLPs/exprstest.jl")
 include("MINLPs/linearlift.jl")
 include("MINLPs/multi.jl")
+include("MINLPs/mult3.jl")
 include("MINLPs/rosenbrock.jl")
-# include("MINLPs/sincos.jl")
-# include("MINLPs/integer.jl")