|
1 | 1 |
|
2 | | - Hipparchus 2.0 RELEASE NOTES |
3 | | - |
4 | | -The Hipparchus team is pleased to announce the release of hipparchus-2.0 |
| 2 | + Hipparchus 2.1 RELEASE NOTES |
5 | 3 |
|
| 4 | +The Hipparchus team is pleased to announce the release of hipparchus-2.1 |
6 | 5 | The Hipparchus library is a library of mathematical components. It started as a fork of Apache Commons Math. |
| 6 | +This is a maintenande release. The main changes are: |
7 | 7 |
|
8 | | -This is a major release. The main changes is the addition of a sequential Gauss-Newton optimizer, |
9 | | -improved relative scheduling of step handlers and events handlers in ODE, |
10 | | -addition of a finish callback in step handlers, replacement of RealFieldElement by CalculusFieldElement interface, |
11 | | -many improvements in Complex to fully implement CalculusFieldElement with correct branch cuts so complex numbers |
12 | | -can be used in many Hipparchus algorithms, addition of FieldComplex, faster simultaneous computation of sinh/cosh. |
13 | | - |
14 | | -Experimental implementation of elliptic functions and integrals for real, complex and general fields have been added. |
15 | | -Beware the elliptic integrals for complex and field complex numbers in the incomplete case are considered experimental |
16 | | -for now, they have known issues (see https://github.com/Hipparchus-Math/hipparchus/issues/151 and |
17 | | -https://github.com/Hipparchus-Math/hipparchus/issues/152). |
18 | | - |
| 8 | + - addition of 3D version of complex function plotter, |
| 9 | + - addition of ConvergenceCheckerOrMultiplexer and ConvergenceCheckerAndMultiplexer, |
| 10 | + - addition of inverse Jacobi elliptic functions, |
| 11 | + - optional modified weights to Akima interpolation, |
| 12 | + - modified Gram-Schmidt basis orthonormalization process, |
| 13 | + - generation of convergents streams in continued fractions, |
| 14 | + - addition of equalsIee754 to points in all topologies to handle NaN coordinates. |
19 | 15 |
|
20 | | -Methods deprecated during the 1.X series have been removed. Numerous bugs were also fixed. |
| 16 | + Various bugs were fixed. |
21 | 17 |
|
22 | 18 | Changes in this version include: |
23 | 19 |
|
24 | 20 | New features: |
25 | | -o issues/153: Added field versions of univariate integrators. |
26 | | -o issues/149: Added complex and field complex univariate integrals (including contour integrals). |
27 | | -o issues/147: Allow retrieving event handler configuration from ODE integrators. |
28 | | -o issues/146: Added a separate finish method in ODE step handlers to be called after last step. |
29 | | -o issues/143: Added ArrayFieldVector.toString(). |
30 | | -o Added sequential Gauss-Newton optimizer. Thanks to Julie Bayard. |
31 | | -o issues/135: Added getRowDimension and getColumnDimension to DecompositionSolver and FieldDecompositionSolver. |
32 | | -o issues/141: Added a way to have a more accurate value of π for fields implementations that require it. |
33 | | -o Added Carlson elliptic integrals (R_F, R_J, R_G, R_D, and R_C) |
34 | | - both for real, CalculusFieldElement, Complex, and FieldComplex. |
35 | | -o issues/124: Changed return types for field elements methods (including Complex): |
36 | | - norm() → double, abs() → field element. |
37 | | -o Added the four Jacobi theta functions θ₁(z|τ), θ₂(z|τ), θ₃(z|τ), and θ₄(z|τ), |
38 | | - both for complex and field complex. |
39 | | -o issues/126: Added field-based complex numbers. |
40 | | -o Added Legendre elliptic integrals (K, K', E, D, F, Π), both complete and incomplete, |
41 | | - both for real, CalculusFieldElement, Complex, and FieldComplex. |
42 | | -o issues/119: Added the twelve Jacobi elliptic functions sn, cn, dn, |
43 | | - cs, ds, ns, dc, nc, sc, nd, sd, and cd, both for primitive double |
44 | | - and fields. |
45 | | -o Added the Ryū algorithm to generate very fast the shortest decimal |
46 | | - representation of a floating point number that maintains round-trip safety. |
| 21 | +o issue/172: Added optional modified weights to avoid overshoots in Akima interpolation |
| 22 | + near constant slopes sub-samples. |
| 23 | +o issue/136: Added ConvergenceCheckerOrMultiplexer and ConvergenceCheckerAndMultiplexer. |
| 24 | +o Added equalsIee754 to S1Point, S2Point, Vector1D, Vector2D and Vector3D |
| 25 | + to have an equality method consistent with IEEE754 concerning NaN coordinates, |
| 26 | + in addition to the regular equals method compliant with Java language design |
| 27 | + choices, including Double.equals (which is different from the double == operator). |
| 28 | +o issue/176: Allow generation of continued fractions convergents as streams of BigFraction/Fraction |
| 29 | + to allow user to apply their own convergence criterion to select one convergent. |
| 30 | +o issue/173: Fixed inconsistency of eigenvalues and eigenvectors between OrderedComplexEigenDecomposition |
| 31 | + and base class ComplexEigenDecomposition. |
47 | 32 |
|
48 | 33 | Fixed Bugs: |
49 | | -o issues/145: Fixed out-of-order scheduling of calls to step handlers and event handlers. |
50 | | -o issues/144: Fixed array index error in field LU decomposition. Thanks to Axel Kramer. |
51 | | -o issues/142: Fixed ulp() implementation for Dfp. |
52 | | -o issues/140: Clarified arguments in the one-argument version of the two-arguments atan2 |
53 | | - function in CalculusFieldElement. |
54 | | -o issues/138: Renamed signum into sign in field classes, and improved implementation for complex numbers. |
55 | | -o issues/131: Fixed consistent implementations between {Field}LUDecomposition classes. |
56 | | -o issues/129: Fixed branch cut on imaginary axis for complex atan. |
57 | | -o issues/123: Added field implementations of ulp function. |
58 | | -o issues/122: Added a way to compute both sinh and cosh at once faster than performing two calls. |
59 | | - This also speeds up all derivatives implementations as well as complex computations. |
60 | | -o issues/121: Removed methods and constants deprecated during the 1.x series. |
61 | | -o issues/67: Replaced RealFieldElement by CalculusFieldElement interface and have Complex |
62 | | - implement it so it can be used in many Hipparchus algorithms (like ODE). |
63 | | - Also replaced RealField{Univariate|Bivariate|Vector|Matrix}Function by |
64 | | - CalculusField{Univariate|Bivariate|Vector|Matrix}Function. |
65 | | -o issues/120: Added event filtering for field ODE. |
66 | | -o issues/118: Use primitive double for initial step and error factors in Field adaptive step size integrators. |
67 | | -o issues/112: Fix interpolation of a pair of ODE states whose times are the same. Such a pair can |
68 | | - appear when a RESET_STATE or RESET_DERIVATIVES event occurs at the integration |
69 | | - stop time. Thanks to andrewsgoetz. |
70 | | -o issues/111: Allow selecting customized threshold for tests in OrderedComplexEigenDecomposition. |
71 | | -o issues/110: Added zero checker predicate for some linear algebra algorithms. |
72 | | -o issues/109: Fixed exception in OrderedComplexEigenDecomposition when eigenvalues are equal. |
73 | | - |
| 34 | +o issue/189: Fixed some wrong branch cuts in inverse Jacobi elliptic functions in the complex case. |
| 35 | +o issue/186: Avoid building an intermediate complex for computing double norm. |
| 36 | +o issue/175: Fixed overflows for incrementors that are allowed to reach Integer.MAX_VALUE. |
| 37 | +o issue/188: Added inverse Jacobi elliptic functions, for both double and field values. |
| 38 | +o issues/187: Fixed wrong set operation with ArcsSet when two BSP trees share |
| 39 | + the same angle as a cut hyperplane. |
| 40 | +o issues/184: Fixed non-bracketing issue when RESET_STATE slightly moves an event at the start |
| 41 | + of a step and another regular event happens in the first half of the same step |
| 42 | +o issue/183: Avoid too many calls to function in BracketingNthOrderBrentSolver when using |
| 43 | + AllowedSolution.ANY_SIDE. |
| 44 | +o issue/178: Added Modified Gram-Schmidt basis orthonormalization process in MatrixUtils |
| 45 | + for double-based and CalculusFieldElement-based list of vectors. |
| 46 | +o issue/177: Fixed issue in Laguerre solver related to different representations of signed zeros. |
| 47 | + |
| 48 | +Changes: |
| 49 | +o Added a 3D version of complex plots. |
74 | 50 |
|
75 | 51 |
|
76 | 52 | For complete information on Hipparchus, including instructions on how to submit bug reports, |
77 | 53 | patches, or suggestions for improvement, see the Hipparchus website: |
78 | 54 |
|
79 | 55 | https://www.hipparchus.org/ |
80 | | - |
81 | | - |
0 commit comments