v0.6.1

IntervalRootFinding v0.6.1

Diff since v0.6.0

Merged pull requests:

• Add iterate method for Root objects (#190) (@edtrelo)
• Add few missing tests (#205) (@Kolaru)
• Add doc, some tests and a minor fix for input and output types in multidimensional cases (#208) (@Kolaru)
• Implement max_iteration and reltol stop criterion (#211) (@Kolaru)
• Update workflows (#214) (@abhro)
• Documentation updates (#215) (@abhro)
• Bump actions/checkout from 2 to 4 (#216) (@dependabot[bot])
• CompatHelper: bump compat for ForwardDiff to 1, (keep existing compat) (#217) (@github-actions[bot])
• Update benchmark and add them to CI (#218) (@Kolaru)
• Add compat for IntervalArithmetic v0.23 and bump version 0.6.1 (#223) (@abhro)

Closed issues:

• Cannot use complex functions (#40)
• Limit the maximum number of intervals (#59)
• Need absolute and relative tolerance (#66)
• Add test with gradient (#87)
• Return semi-infinite intervals when cannot guarantee roots (#96)
• Make a RootProblem type (#109)
• Allow keywords arguments for roots (#116)
• No method roots(f, v::Vector{IntervalBox}) (#117)
• Can't just specify a tolerance on master (#136)
• Incorrect root of function involving max (#146)
• How to show there is no roots on unbounded domain? (#149)
• Throw an error when input and output dimension mismatch (#150)
• Graceful handling of infinity (#161)
• Specifying a maximum number of bisections (#178)
• Allow to unpack roots objects (#180)
• Empty interval results on discontinuous functions (#184)
• Multithreading causes random errors (#189)
• Missing import from IntervalArithmetic? (#200)
• Update versions of dependencies (#201)
• ERROR: MethodError: no method matching contains_zero(::Float64) (#202)
• Where are docs for 0.6 / what are all the breaking changes? (#207)
• Usage of \times and .. symbols in documentation (#212)
• Kindly provide an example for newer syntax (#213)

v0.6.0

IntervalRootFinding v0.6.0

Diff since v0.5.10

Main change

• Compatibility with IntervalArithmetic v0.22

Breaking changes

• Decorated intervals must be used (which is the default Interval in the IntervalArithmetic v0.22)

• The signature of the roots function changed to roots(f::Function, X::Union{Interval, AbstractVector} ; kwargs...), with the following consequences

  • Manual derivatives and contractors must now always be passed as keyword arguments. This greatly simplify the internal logic of the roots function.
  • No more IntervalBox. Multidimensional problem are specified by returning a vector of intervals, and giving a vector of intervals as initial search region.
  • Normal vectors of intervals are supported. They are significantly (3x) time slower for a simple 2D problem than SVector, but more convenient.

• Features of the packages that were undocumented are now unsupported (e.g. Slopes and quadratic equations). If you were using them, please open an issue.

New feature

• Standard vectors of intervals are supported. They are significantly (3x) time slower for a simple 2D problem than SVector, but more convenient.

Merged pull requests:

• Bump version for Polynomials v3 (#183) (@hurak)
• Use branch and prune as backend for the search (#187) (@Kolaru)
• Refactor contractors (#188) (@Kolaru)
• update docs workflow (#194) (@lucaferranti)
• Compatibility with IntervalArithmetic v0.22 (#203) (@Kolaru)

Closed issues:

• Move lexcmp on Intervals to IntervalArithmetic.jl (#4)
• Use queue instead of stack (#22)
• Change to work with AbstractInterval (#23)
• Bisection of atomic interval (#26)
• Add debug mode (#30)
• Remove old code (#41)
• Clean up code (#52)
• Atomic intervals should be treated specially (#58)
• roots cannot handle functions that return intervals (#77)
• Example 5.5 from Smiley and Chun (2001) solved inconsistently (#86)
• Newton fails with abs (#98)
• Support function returning IntervalBox and only use them internally (#107)
• Do not export Contractor (#128)
• Factor out branch and bound into new package (#130)
• Allow roots(f, Xs) for a vector Xs (#157)
• Extra roots bug for large intervals (#162)
• Incorrect unique zero for 0/x + x and similar. (#181)
• ForwardDiff cannot guarantee valid enclosures of derivatives (#182)
• Update to make compatible with IntervalArithmetic master (#191)
• Doc badge does not point to the package doc (#193)
• Unexpected answer to trigonometric polynomial (#195)
• tan tangent function missing roots (#196)
• ERROR: SingularException(4)....what happen? (#198)
• ERROR: SingularException(4) (#199)

v0.5.10

IntervalRootFinding v0.5.10

Diff since v0.5.9

Merged pull requests:

• CompatHelper: bump compat for IntervalArithmetic to 0.20, (keep existing compat) (#179) (@github-actions[bot])

v0.5.9

IntervalRootFinding v0.5.9

Diff since v0.5.8

Merged pull requests:

• CompatHelper: bump compat for "IntervalArithmetic" to "0.19" (#177) (@github-actions[bot])

v0.5.8

IntervalRootFinding v0.5.8

Diff since v0.5.7

Closed issues:

• Export interval type (#163)
• Tax Title IV-D

Waste;

(#172)

• Acces to elements / sort (#174)

Merged pull requests:

• export interval definition for convenience (#176) (@Datseris)

v0.5.7

IntervalRootFinding v0.5.7

Diff since v0.5.6

Closed issues:

• bump version number of Polynomials (#170)

Merged pull requests:

• updated compat for IntervalArithmetic (#171) (@lucaferranti)

v0.5.6

IntervalRootFinding v0.5.6

Diff since v0.5.5

Closed issues:

• Add root-status "root(s)" ? (#164)

Merged pull requests:

• updated workflow to gha (#165) (@lucaferranti)

v0.5.5

IntervalRootFinding v0.5.5

Diff since v0.5.4

Merged pull requests:

• Update Project.toml (#159) (@mforets)

v0.5.4

IntervalRootFinding v0.5.4

Diff since v0.5.3

Closed issues:

• Add hard test cases (#44)
• roots hangs with double roots (#74)
• roots should not return "unique" for multiple roots (#75)
• Newton fails for dummy example with continuous solutions (#101)
• Setting the tolerance too low for BigFloat makes roots disappear (#129)
• Released version is broken (#154)

Merged pull requests:

• Install TagBot as a GitHub Action (#145) (@JuliaTagBot)
• Allow using IntervalArithmetic.jl v0.17 (#152) (@lbenet)
• Reenable and fix example 5.5 from Smiley et al. (#153) (@gwater)
• Bump Polynomials version and install CompatHelper (#155) (@dpsanders)
• Make bisection test more expressive (#156) (@gwater)

v0.5.3

Allow using IntervalArithmetic.jl v0.17 (#152)

* Allow using IntervalArithmetic.jl v0.17

* Add tests for Julia 1.4, and fix Polynomials version