Coefficient Strengthening #14

Anhtu07 · 2020-12-31T02:05:22Z

Add coefficient strengthening for presolve

! Empty rows/columns are now stored with empty arrays for nzind and nzval

-Implement and create tests for coef strengthening -Ignore ranged constraints (i.e., d <= a'x <= b)

-Adding more test on coefficient strengthening -CS now removes 0 coefficient, and empty column is now respresented as []

-Add coef strengthening to presolve.jl -Add testset

test/runtests.jl

mtanneau · 2020-12-31T02:49:25Z

src/mip/coefficient_strengthening.jl

@@ -0,0 +1,99 @@
+function maximal_activity(row::RowOrCol{T}, j::Int, ucol::Vector{T}, lcol::Vector{T})::T where {T}


What is this function doing exactly? Are you trying to compute the maximum activity of a given row?

If so, your implementation is not correct: there may be some variables that have been removed from the problem.
You need to explicitly check this with ps.colflag.

I suggest you use the code that's currently in lp/forcing_row:

# Implied row bounds row = ps.pb0.arows[i] l_ = u_ = zero(T) for (j, aij) in zip(row.nzind, row.nzval) ps.colflag[j] || continue if aij < zero(T) l_ += aij * ps.ucol[j] u_ += aij * ps.lcol[j] else l_ += aij * ps.lcol[j] u_ += aij * ps.ucol[j] end isfinite(l_) || isfinite(u_) || break # infinite bounds end

You need access to ps (the PresolveData struct), and the row row you're interested in (or its index i).
At the end of this loop, you get l_ and u_ which are the lower and upper activity bounds on row i.
I you want to skip variable j0 in the computation, just update the check to

(ps.colflag[j] && j != j0) || continue

To @mtanneau's point, a short docstring above the method definition would be helpful to understand what this function does.

It would be good to move the code snippet from forcing_row to a separate util.jl file.

src/mip/coefficient_strengthening.jl

src/presolve.jl

mtanneau · 2020-12-31T02:59:01Z

src/presolve.jl

@@ -414,6 +415,9 @@ function presolve!(ps::PresolveData{T}) where {T}
    # Round the bounds of integer variables are integers.
    round_integer_bounds!(ps)

+    # Coeficient strengthening


Suggested change

# Coeficient strengthening

# Coefficient strengthening

mtanneau · 2020-12-31T02:59:56Z

src/presolve.jl

+Perform coefficient strengthening for integer/binary variables
+on every constraints.
+
+Called once at the very beginning of the presolve procedure.


Suggested change

Called once at the very beginning of the presolve procedure.

Called once at the beginning of the presolve procedure.

mtanneau · 2020-12-31T03:03:01Z

src/mip/coefficient_strengthening.jl

+function single_row_strengthening(row::RowOrCol{T}, b::T, m::Int, j::Int, ucol::Vector{T}, lcol::Vector{T}) where {T}
+    # perform coef strengthening for one constraints of the from a'x < b
+    # and return new a' and b'
+    flag = false # wether there is an update


Two spaces after the end of the line + typo

Suggested change

flag = false # wether there is an update

flag = false # whether there is an update

mtanneau · 2020-12-31T03:03:12Z

src/mip/coefficient_strengthening.jl

+
+        m = findfirst(isequal(j), row.nzind)
+        if m == nothing || (lrow > -inf && urow < inf)
+            continue #skipping ranged constraints and


Suggested change

continue #skipping ranged constraints and

continue # skipping ranged constraints and

mtanneau · 2020-12-31T03:07:18Z

src/mip/coefficient_strengthening.jl

+
+                # update collumns of A in problem data
+                # and update the sparse matrix in the case where the new coefficient is close to 0
+                if abs(new_coef) <= eps(T)


We should use a drop tolerance here.
cc @joehuchette

Agreed. The eps(T) should be a configurable tolerance in PresolveOptions. Also, this should likely be a separate DropSmallCoefficients Rule.

mtanneau · 2020-12-31T03:27:17Z

There is some work to do before this can be merged.
Here are the main points I want to improve:

Correctly handle variables/rows that have been deactivated.
I understand that, currently, the routine is only called at the beginning of presolve, but this may change.
If it's absolutely needed, introduce two new parameters in PresolveOptions:
- drop tolerance for rounding small coefficients to zero
- "infinity" threshold for rounding large values to infinity
Performance improvements.
I'm not looking for lightning speeds, but I've seen a lot of room for improvements (both time and memory-wise).
The goal is that the routine can be run in O(nz), where nz is the number of non-zero coefficients in A.
- Go row-by-row instead of column-by-column:
  1. For each row, compute its activity bounds.
    At the same time, check that at least one variable is integer, otherwise move to next row.
  2. For each variable that appears in the row, perform coefficient strengthening, if applicable. This can be done in-place.
    If a coefficient is strengthened, I think it's possible to immediately update the row's activity bounds, before moving to the next variable.
  3. If some coefficients do end up being rounded to zero, the non-zero counters for said row and column must be updated
- Avoid (deep) copies. We shouldn't need any.
Make the code work for range constraints.
All constraints are stored as range constraints, which simply denote a'x <= u and -a'x <= -l.
Thus, it should be possible to treat both cases in a single pass on each row. I'm 99.9% sure there won't be any conflict.
Add some documentation.

[EDIT: dropping range constraints]

joehuchette · 2020-12-31T14:24:48Z

src/mip/coefficient_strengthening.jl

@@ -0,0 +1,99 @@
+function maximal_activity(row::RowOrCol{T}, j::Int, ucol::Vector{T}, lcol::Vector{T})::T where {T}


To @mtanneau's point, a short docstring above the method definition would be helpful to understand what this function does.

joehuchette · 2020-12-31T14:26:20Z

src/mip/coefficient_strengthening.jl

@@ -0,0 +1,99 @@
+function maximal_activity(row::RowOrCol{T}, j::Int, ucol::Vector{T}, lcol::Vector{T})::T where {T}


It would be good to move the code snippet from forcing_row to a separate util.jl file.

joehuchette · 2020-12-31T14:27:03Z

src/mip/coefficient_strengthening.jl

+end
+
+function single_row_strengthening(row::RowOrCol{T}, b::T, m::Int, j::Int, ucol::Vector{T}, lcol::Vector{T}) where {T}
+    # perform coef strengthening for one constraints of the from a'x < b


Suggested change

# perform coef strengthening for one constraints of the from a'x < b

# perform coefficient strengthening for one constraint of the form a'x \leq b

What is the mathematical definition for the coefficient strengthening you are doing? Can you include it in the docstring?

joehuchette · 2020-12-31T14:28:31Z

src/mip/coefficient_strengthening.jl

+    a = deepcopy(row.nzval)
+    if a[m] > 0
+        d = b - maximal_activity(row, j, ucol, lcol) - a[m]*(ucol[j]-1)
+        if a[m] >= d && d > 0


Small suggestion: These can be chained as a[m] >= d > 0

src/mip/coefficient_strengthening.jl

joehuchette · 2020-12-31T14:38:25Z

src/mip/coefficient_strengthening.jl

+                row.nzval = -a
+                ps.lrow[i] = -b
+
+                # update collumns of A in problem data


Suggested change

# update collumns of A in problem data

# update columns of A in problem data

joehuchette · 2020-12-31T14:41:07Z

test/mip/coefficient_strengthening.jl

+    C = T[1, 1, 1, 1]
+    lc = T[0, 1, 0, 2]
+    uc = T[1, 5//2, 1, 3]
+    lr = T[-1e30, -1e30]


Update for new Inf standard

joehuchette · 2020-12-31T14:41:20Z

test/mip/coefficient_strengthening.jl

+
+
+    @test ps.urow == T[5, 9//2]
+    @test ps.lrow == T[-1e30, -1e30]


joehuchette · 2020-12-31T14:41:25Z

test/mip/coefficient_strengthening.jl

+   C = T[1, 1, 1, 1]
+   lc = T[0, 0, 0, -3]
+   uc = T[2, 1, 2, 10^30]
+   lr = T[-1e30, -1e30]


joehuchette · 2020-12-31T14:41:35Z

test/mip/coefficient_strengthening.jl

+    lc = T[-4, -3//2, 9, 0, -1]
+    uc = T[2, 1, 15, 1, 1]
+    lr = T[2, -1e30]
+    ur = T[1e30, 13/3]


joehuchette · 2020-12-31T14:48:38Z

My concern about ranged constraints is that if you apply the strengthening to both sides you will end up with distinct linear pieces, which would require us adding extra constraints to our representation. Is this not the case? If it is, is it worth the added complexity?

mtanneau · 2020-12-31T15:03:50Z

My concern about ranged constraints is that if you apply the strengthening to both sides you will end up with distinct linear pieces, which would require us adding extra constraints to our representation. Is this not the case? If it is, is it worth the added complexity?

Good point :)
For a row l <= a'x <= u, I guess this also raises the question of whether the coefficient strengthening w.r.t a'x <= l is valid w.r.t l <= a'x.

If the answer is no, we can skip range constraints.
If the answer is yes, we have 2 options:
1. Choose either side and do the coefficient strengthening w.r.t that side only.
2. If there's no conflict, try both strengthenings and keep the best (where "best" is TBD)

If it's too much hassle at this point we can drop it.

joehuchette · 2020-12-31T16:49:23Z

@Anhtu07: Thoughts?

My initial (offline) suggestion was to punt completely on range constraints for the time being; I hadn't considered the "yes" branch. It's worth thinking about.

Anhtu07 · 2020-12-31T18:19:44Z

I think there will be conflict most of the time when using CS on ranged constraints. For example, consider
3 <= 2x + 3y <= 6, 0<= x <=1 and y is integral, y<=2, using CS w.r.t 2x + 3y <=6 will get 2x + 2y <= 4, but 3<= 2x + 2y isn't valid.
Intuitively, when using CS, we decrease/increase some coefficient toward 0, and thus the lower/upper need to be updated as well

mtanneau · 2020-12-31T19:15:30Z

there will be conflict most of the time when using CS on ranged constraints

All right, I'm convinced. Let's skip range-constraints.

- using Inf instead of 1e30

codecov-io · 2021-01-01T23:52:42Z

Codecov Report

Merging #14 (3e7fdc3) into master (c317b25) will increase coverage by 5.78%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #14      +/-   ##
==========================================
+ Coverage   46.83%   52.62%   +5.78%     
==========================================
  Files          14       15       +1     
  Lines         901     1011     +110     
==========================================
+ Hits          422      532     +110     
  Misses        479      479

Impacted Files	Coverage Δ
src/mip/coefficient_strengthening.jl	`100.00% <100.00%> (ø)`
src/presolve.jl	`47.25% <100.00%> (+1.83%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c317b25...72cb1ad. Read the comment docs.

mtanneau · 2021-01-02T02:50:40Z

src/mip/coefficient_strengthening.jl

@@ -52,9 +52,9 @@ function coefficient_strengthening!(ps::PresolveData{T}, j::Int, inf::T = T(1e30
        urow = ps.urow[i]

        m = findfirst(isequal(j), row.nzind)
-        if m == nothing || (lrow > -inf && urow < inf)
+        if m == nothing || (lrow > -Inf && urow < Inf)


just to ensure type-stability

Suggested change

if m == nothing || (lrow > -Inf && urow < Inf)

if m == nothing || (lrow > T(-Inf) && urow < T(Inf))

mtanneau · 2021-01-02T02:50:48Z

src/mip/coefficient_strengthening.jl

            continue #skipping ranged constraints and
-        elseif urow < inf
+        elseif urow < Inf


Suggested change

elseif urow < Inf

elseif urow < T(Inf)

mtanneau · 2021-01-02T02:50:58Z

src/mip/coefficient_strengthening.jl

@@ -73,7 +73,7 @@ function coefficient_strengthening!(ps::PresolveData{T}, j::Int, inf::T = T(1e30
                    ps.pb0.acols[j].nzval[k] = new_coef
                end
            end
-        elseif lrow > -inf
+        elseif lrow > -Inf


Suggested change

elseif lrow > -Inf

elseif lrow > T(-Inf)

src/mip/coefficient_strengthening.jl

- remove deepcopy - perform coefficient strengthening row by row instead of col by col

- no longer using findfirst, which may cost nzlog(nz) worst-case time complexity

- Update way in which the maximal activity of a row except variable j is compute

- skip row with no integer variable - set rowflag/colflag - update #nonzeros in each row/col - add docs

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

src/mip/coefficient_strengthening.jl

- update all problem's parameters in the same block

joehuchette · 2021-01-23T18:19:31Z

src/mip/coefficient_strengthening.jl

+Perform coefficient strengthening on each row and on each integer variables
+
+In particular, for the i-th constraints of the form
+aᵢ x ⩽ bᵢ and let xⱼ be a integer/binary variable


End sentence with a period.

joehuchette · 2021-01-23T18:21:05Z

src/mip/coefficient_strengthening.jl

+aᵢⱼ ≔ aᵢⱼ - d
+bᵢ ≔ bᵢ - duⱼ
+Also perform coefficient strengthening on constraints of the form aᵢ x ⩾ cᵢ
+with similar update rule


End with a period

joehuchette · 2021-01-23T18:21:55Z

src/mip/coefficient_strengthening.jl

+Also perform coefficient strengthening on constraints of the form aᵢ x ⩾ cᵢ
+with similar update rule
+
+Skip ranged constraints, i.e, cᵢ ⩽ aᵢ x ⩽ bᵢ


Make a complete sentence (add "We" at beginning, end with a period).

joehuchette · 2021-01-23T18:22:01Z

src/mip/coefficient_strengthening.jl

+
+Skip ranged constraints, i.e, cᵢ ⩽ aᵢ x ⩽ bᵢ
+
+The new coefficients are stored by updating ps.pb0


joehuchette · 2021-01-23T18:23:21Z

src/mip/coefficient_strengthening.jl

+where Mᵢⱼ is maximal activity of the i-th row without considering xⱼ
+and uⱼ is upper bound of xⱼ, then
+aᵢⱼ ≔ aᵢⱼ - d
+bᵢ ≔ bᵢ - duⱼ


This currently reads as an identity; I would rephrase to make clear that this is the transformation you apply.

Suggested change

bᵢ ≔ bᵢ - duⱼ

and uⱼ is upper bound of xⱼ, then we transform

aᵢⱼ <- aᵢⱼ - d

bᵢ <- bᵢ - duⱼ

joehuchette · 2021-01-23T18:33:10Z

src/mip/coefficient_strengthening.jl

+    return new_coef, new_bound
+end
+
+function lowerbound_strengthening!(ps::PresolveData{T}, i::Int, j_index::Int, j::Int, min_act) where {T}


Does this actually modify ps? If not, remove the !. (And similarly for upperbound)

joehuchette · 2021-01-23T18:34:18Z

src/mip/coefficient_strengthening.jl

+                    end
+                    #update sup
+                    if coef > 0
+                        sup = sup - (coef - new_coef) * ps.ucol[j]


Suggested change

sup = sup - (coef - new_coef) * ps.ucol[j]

sup -= (coef - new_coef) * ps.ucol[j]

joehuchette · 2021-01-23T18:34:33Z

src/mip/coefficient_strengthening.jl

+                    if coef > 0
+                        sup = sup - (coef - new_coef) * ps.ucol[j]
+                    else
+                        sup = sup - (coef - new_coef) * ps.lcol[j]


Suggested change

sup = sup - (coef - new_coef) * ps.lcol[j]

sup -= (coef - new_coef) * ps.lcol[j]

joehuchette · 2021-01-23T18:36:11Z

src/mip/coefficient_strengthening.jl

+                    end
+                    #update inf
+                    if coef > 0
+                        inf = inf - (coef - new_coef) * ps.lcol[j]


Suggested change

inf = inf - (coef - new_coef) * ps.lcol[j]

inf -= (coef - new_coef) * ps.lcol[j]

joehuchette · 2021-01-23T18:36:19Z

src/mip/coefficient_strengthening.jl

+                    if coef > 0
+                        inf = inf - (coef - new_coef) * ps.lcol[j]
+                    else
+                        inf = inf - (coef - new_coef) * ps.ucol[j]


Suggested change

inf = inf - (coef - new_coef) * ps.ucol[j]

inf -= (coef - new_coef) * ps.ucol[j]

- add one more test - edit docs (period and complete sentence) - using += and -=

joehuchette · 2021-01-28T00:14:10Z

src/mip/coefficient_strengthening.jl

+Perform coefficient strengthening on each row and on each integer variables
+
+In particular, for the i-th constraints of the form
+aᵢ x ⩽ bᵢ and let xⱼ be a integer/binary variable


joehuchette · 2021-01-28T00:14:32Z

src/mip/coefficient_strengthening.jl

+If a column is reduced to 0 after this step, it is set to be inactive
+
+"""
+function maximal_activity(row::Row{T}, ucol::Vector{T}, lcol::Vector{T})::T where {T}


joehuchette · 2021-01-28T00:14:38Z

src/mip/coefficient_strengthening.jl

        end
    end
-    sup = ucol[nonzero_index]'pos_coef + lcol[nonzero_index]'neg_coef
    return T(sup)


joehuchette · 2021-01-28T00:14:45Z

src/mip/coefficient_strengthening.jl

-            a[m] = a[m] - d
-            b = b - d*ucol[j]
-            flag = true
+function minimal_activity(row::Row{T}, ucol::Vector{T}, lcol::Vector{T})::T where {T}


joehuchette · 2021-01-28T00:14:58Z

src/mip/coefficient_strengthening.jl

+        end
+    end


joehuchette · 2021-01-28T00:15:42Z

src/mip/coefficient_strengthening.jl

+end
+
+function zero_coefficient_strengthening!(ps::PresolveData{T}) where {T}
+    # perform coefficient stregthening but if there is a coefficient is reduced to 0


joehuchette · 2021-01-28T00:16:05Z

src/mip/coefficient_strengthening.jl

+
+function zero_coefficient_strengthening!(ps::PresolveData{T}) where {T}
+    # perform coefficient stregthening but if there is a coefficient is reduced to 0
+    # it is still kept in the ps.pb0.arows and ps.pb0.acols


joehuchette · 2021-01-28T00:16:13Z

src/mip/coefficient_strengthening.jl

+        sup = maximal_activity(row, ps.ucol, ps.lcol)
+        inf = minimal_activity(row, ps.ucol, ps.lcol)


joehuchette · 2021-01-28T00:16:19Z

src/presolve.jl

@@ -414,6 +415,9 @@ function presolve!(ps::PresolveData{T}) where {T}
    # Round the bounds of integer variables are integers.
    round_integer_bounds!(ps)

+    # Coeficient strengthening


joehuchette · 2021-01-28T00:16:31Z

src/presolve.jl

+Perform coefficient strengthening for integer/binary variables
+on every constraints.
+
+Called once at the very beginning of the presolve procedure.


joehuchette · 2021-01-28T00:18:45Z

@Anhtu07 There are some outstanding comments (I think) that I bumped. You can see them if you click the "Files Changed" tab up top.

mtanneau

One more thing: check whether rows/columns are active or not

src/mip/coefficient_strengthening.jl

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

src/mip/coefficient_strengthening.jl

joehuchette · 2021-01-31T14:02:38Z

src/presolve.jl

+    coefficient_strengthening!(ps::PresolveData)
+
+Perform coefficient strengthening for integer/binary variables
+on every constraints.


Suggested change

on every constraints.

in every constraint.

joehuchette · 2021-01-31T14:04:15Z

src/mip/coefficient_strengthening.jl

+    return new_coef, new_bound
+end
+
+function zero_coefficient_strengthening!(ps::PresolveData{T}) where {T}


I'm not sure about this naming. It sounds like you're only applying coefficient strengthening on zero coefficients (maybe?). I would just call this _coefficient_strengthenening!.

joehuchette · 2021-01-31T14:19:51Z

src/mip/coefficient_strengthening.jl

@@ -0,0 +1,178 @@
+"""


The location of this docstring is a bit odd--usually it would go immediately before (no extra newline) before a function it's describing. Here, it's above maximal_activity, but that's just really a helper function.

joehuchette · 2021-01-31T14:21:06Z

src/mip/coefficient_strengthening.jl

+If a row/column is reduced to 0 after this step, it will be set to be inactive.
+"""
+
+function maximal_activity(ps::PresolveData{T}, i::Int)::T where {T}


Consider moving maximal_activity and minimal_activity to util.jl.

mtanneau

I think we're reaching the end!
@Anhtu07 you will be the record holder for the longest PR :)

I have checked all the source code. Except for the removal of empty rows/columns, everything looks good. I have not checked the underlying math.
I have also not checked the correctness of tests, I trust you know what you're doing.

test/mip/coefficient_strengthening.jl

src/presolve.jl

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

Co-authored-by: Joey Huchette <joehuchette@gmail.com>

- move maximal_activity/ minimal_activity to util - remove coefficient_strengthening! from presolve.jl - deleting spaces

Anhtu07 · 2021-01-31T21:20:25Z

@mtanneau Thanks for letting me know, I will put that on my CV :) Trust me, future PR will not take as long

you can review it one more time now.

mtanneau

LGTM

Given the large number of commits in this PR, I'd prefer we squash and merge, and clean the commit history when doing so.

joehuchette · 2021-02-01T01:45:59Z

Thanks for your hard work on this one, @Anhtu07 !

Anhtu07 added 13 commits December 20, 2020 19:34

test on julia 1.4

72c5e67

Merge branch 'newfeature'

f807aa3

Merge remote-tracking branch 'mtanneau/master' into newfeature

46c0ec4

Merge branch 'newfeature'

61e1e4f

Merge remote-tracking branch 'mtanneau/master'

c683ca1

Merge branch 'master' into newfeature

c83a339

First commit for coefficient strengthening

7a8afd3

-Implement and create tests for coef strengthening -Ignore ranged constraints (i.e., d <= a'x <= b)

Add test3

a41ddfd

-Adding more test on coefficient strengthening -CS now removes 0 coefficient, and empty column is now respresented as []

Add coef strengthening to presolve.jl

89bb756

-Add coef strengthening to presolve.jl -Add testset

Merge branch 'newfeature'

ceaf8cb

remove julia 1.4

6387ed5

Merge remote-tracking branch 'mtanneau/master'

4193cef

Merge branch 'master' into newfeature

394e729

mtanneau requested changes Dec 31, 2020

View reviewed changes

joehuchette requested changes Dec 31, 2020

View reviewed changes

remove infinity

531ac5e

- using Inf instead of 1e30

mtanneau reviewed Jan 2, 2021

View reviewed changes

BochuanBob reviewed Jan 2, 2021

View reviewed changes

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

Anhtu07 added 5 commits January 6, 2021 14:49

Merge branch 'master' into newfeature

f429307

row by row iteration

e5e945c

- remove deepcopy - perform coefficient strengthening row by row instead of col by col

update coef_strengthening

12f54be

- no longer using findfirst, which may cost nzlog(nz) worst-case time complexity

O(nz) update

4038192

- Update way in which the maximal activity of a row except variable j is compute

Update

d2b4cce

- skip row with no integer variable - set rowflag/colflag - update #nonzeros in each row/col - add docs

Update src/mip/coefficient_strengthening.jl

fb5e4dd

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

mtanneau reviewed Jan 22, 2021

View reviewed changes

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

Anhtu07 added 5 commits January 22, 2021 16:30

Merge branch 'master' into newfeature

9f8a6b4

consistent update

e964e7b

- update all problem's parameters in the same block

Merge branch 'newfeature'

1707cd8

using isfinite

a2faa0a

minor fix

1a29210

joehuchette requested changes Jan 23, 2021

View reviewed changes

Anhtu07 added 2 commits January 23, 2021 14:41

Add test

3eb946e

- add one more test - edit docs (period and complete sentence) - using += and -=

fix

6eb7c75

joehuchette requested changes Jan 28, 2021

View reviewed changes

mtanneau requested changes Jan 28, 2021

View reviewed changes

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

src/mip/coefficient_strengthening.jl Show resolved Hide resolved

Anhtu07 added 3 commits January 30, 2021 16:36

Merge branch 'master' into newfeature

3a1d952

skip deactivated rows/cols

0b4b83e

fix bumped comments

1ea9b79

mtanneau reviewed Jan 30, 2021

View reviewed changes

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

mtanneau reviewed Jan 30, 2021

View reviewed changes

src/mip/coefficient_strengthening.jl Outdated Show resolved Hide resolved

Anhtu07 and others added 2 commits January 30, 2021 17:55

Update src/mip/coefficient_strengthening.jl

3e7fdc3

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

Update src/mip/coefficient_strengthening.jl

72cb1ad

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

joehuchette approved these changes Jan 31, 2021

View reviewed changes

mtanneau requested changes Jan 31, 2021

View reviewed changes

Anhtu07 and others added 3 commits January 31, 2021 09:22

Update test/mip/coefficient_strengthening.jl

021f501

Co-authored-by: mtanneau <9593025+mtanneau@users.noreply.github.com>

Update src/mip/coefficient_strengthening.jl

8ae6572

Co-authored-by: Joey Huchette <joehuchette@gmail.com>

Update

92d07a6

- move maximal_activity/ minimal_activity to util - remove coefficient_strengthening! from presolve.jl - deleting spaces

mtanneau approved these changes Jan 31, 2021

View reviewed changes

joehuchette merged commit e29763f into mtanneau:master Feb 1, 2021

mtanneau mentioned this pull request Feb 1, 2021

MIP presolve reductions roadmap #9

Open

		@@ -0,0 +1,99 @@
		function maximal_activity(row::RowOrCol{T}, j::Int, ucol::Vector{T}, lcol::Vector{T})::T where {T}

	Called once at the very beginning of the presolve procedure.
	Called once at the beginning of the presolve procedure.

	flag = false # wether there is an update
	flag = false # whether there is an update

	continue #skipping ranged constraints and
	continue # skipping ranged constraints and

	# perform coef strengthening for one constraints of the from a'x < b
	# perform coefficient strengthening for one constraint of the form a'x \leq b

	# update collumns of A in problem data
	# update columns of A in problem data

	if m == nothing \|\| (lrow > -Inf && urow < Inf)
	if m == nothing \|\| (lrow > T(-Inf) && urow < T(Inf))


		Skip ranged constraints, i.e, cᵢ ⩽ aᵢ x ⩽ bᵢ

		The new coefficients are stored by updating ps.pb0

-bᵢ ≔ bᵢ - duⱼ
+and uⱼ is upper bound of xⱼ, then we transform
+aᵢⱼ <- aᵢⱼ - d
+bᵢ <- bᵢ - duⱼ

	sup = sup - (coef - new_coef) * ps.ucol[j]
	sup -= (coef - new_coef) * ps.ucol[j]

	sup = sup - (coef - new_coef) * ps.lcol[j]
	sup -= (coef - new_coef) * ps.lcol[j]

	inf = inf - (coef - new_coef) * ps.lcol[j]
	inf -= (coef - new_coef) * ps.lcol[j]

	inf = inf - (coef - new_coef) * ps.ucol[j]
	inf -= (coef - new_coef) * ps.ucol[j]

		sup = maximal_activity(row, ps.ucol, ps.lcol)
		inf = minimal_activity(row, ps.ucol, ps.lcol)

Coefficient Strengthening #14

Coefficient Strengthening #14

Conversation

Anhtu07 commented Dec 31, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mtanneau commented Dec 31, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

joehuchette commented Dec 31, 2020

mtanneau commented Dec 31, 2020

joehuchette commented Dec 31, 2020

Anhtu07 commented Dec 31, 2020

mtanneau commented Dec 31, 2020

codecov-io commented Jan 1, 2021 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

joehuchette commented Jan 28, 2021

mtanneau left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mtanneau left a comment

Choose a reason for hiding this comment

Anhtu07 commented Jan 31, 2021

mtanneau left a comment • edited Loading

Choose a reason for hiding this comment

joehuchette commented Feb 1, 2021

mtanneau commented Dec 31, 2020 •

edited

Loading

codecov-io commented Jan 1, 2021 •

edited

Loading

mtanneau left a comment •

edited

Loading