Skip to content

Commit

Permalink
renew project
Browse files Browse the repository at this point in the history
  • Loading branch information
@linjing-lab
linjing-lab committed Nov 10, 2022
1 parent dc57a33 commit c78f7f7
Show file tree
Hide file tree
Showing 25 changed files with 345 additions and 350 deletions.
50 changes: 25 additions & 25 deletions README.md
View file

Large diffs are not rendered by default.

50 changes: 25 additions & 25 deletions README_en.md
View file

Large diffs are not rendered by default.

22 changes: 8 additions & 14 deletions examples/doc/_constrain.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,8 @@ oc.equal.[函数名]([目标函数], [参数表], [等式约束表], [初始迭

| 方法头 | 解释 |
| ----------------------------------------------------------------------------------------------------------------------------------------------------- | --------- |
| penalty_quadratice(funcs: FuncArray, args: FuncArray, cons: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", sigma: Optional[float]=10, p: Optional[float]=2, epsilon: Optional[float]=1e-4, k: Optional[int]=0) -> OutputType | 增加二次罚项 |
| lagrange_augmentede(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", lamk: Optional[float]=6, sigma: Optional[float]=10, p: Optional[float]=2, etak: Optional[float]=1e-4, epsilon: Optional[float]=1e-6, k: Optional[int]=0) -> OutputType | 增广拉格朗日乘子法 |
| penalty_quadratice(funcs: FuncArray, args: FuncArray, cons: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", sigma: float=10, p: float=2, epsilon: float=1e-4, k: int=0) -> OutputType | 增加二次罚项 |
| lagrange_augmentede(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", lamk: float=6, sigma: float=10, p: float=2, etak: float=1e-4, epsilon: float=1e-6, k: int=0) -> OutputType | 增广拉格朗日乘子法 |


```python
Expand Down Expand Up @@ -60,9 +60,9 @@ oc.unequal.[函数名]([目标函数], [参数表], [不等式约束表], [初

| 方法头 | 解释 |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------- |
| penalty_quadraticu(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", sigma: Optional[float]=10, p: Optional[float]=0.4, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 增加二次罚项 |
| penalty_interior_fraction(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", sigma: Optional[float]=12, p: Optional[float]=0.6, epsilon: Optional[float]=1e-6, k: Optional[int]=0) -> OutputType | 增加分式函数罚项 |
| lagrange_augmentedu(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", muk: Optional[float]=10, sigma: Optional[float]=8, alpha: Optional[float]=0.2, beta: Optional[float]=0.7, p: Optional[float]=2, eta: Optional[float]=1e-1, epsilon: Optional[float]=1e-4, k: Optional[int]=0) -> OutputType | 增广拉格朗日乘子法 |
| penalty_quadraticu(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", sigma: float=10, p: float=0.4, epsilon: float=1e-10, k: int=0) -> OutputType | 增加二次罚项 |
| penalty_interior_fraction(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", sigma: float=12, p: float=0.6, epsilon: float=1e-6, k: int=0) -> OutputType | 增加分式函数罚项 |
| lagrange_augmentedu(funcs: FuncArray, args: ArgArray, cons: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", muk: float=10, sigma: float=8, alpha: float=0.2, beta: float=0.7, p: float=2, eta: float=1e-1, epsilon: float=1e-4, k: int=0) -> OutputType | 增广拉格朗日乘子法 |


```python
Expand Down Expand Up @@ -92,9 +92,9 @@ oc.mixequal.[函数名]([目标函数], [参数表], [等式约束表], [不等

| 方法头 | 解释 |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- |
| penalty_quadraticm(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", sigma: Optional[float]=10, p: Optional[float]=0.6, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 增加二次罚项 |
| penalty_L1(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", sigma: Optional[float]=1, p: Optional[float]=0.6, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | L1精确罚函数法 |
| lagrange_augmentedm(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="gradient_descent", lamk: Optional[float]=6, muk: Optional[float]=10, sigma: Optional[float]=8, alpha: Optional[float]=0.5, beta: Optional[float]=0.7, p: Optional[float]=2, eta: Optional[float]=1e-3, epsilon: Optional[float]=1e-4, k: Optional[int]=0) -> OutputType | 增广拉格朗日乘子法 |
| penalty_quadraticm(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", sigma: float=10, p: float=0.6, epsilon: float=1e-10, k: int=0) -> OutputType | 增加二次罚项 |
| penalty_L1(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", sigma: float=1, p: float=0.6, epsilon: float=1e-10, k: int=0) -> OutputType | L1精确罚函数法 |
| lagrange_augmentedm(funcs: FuncArray, args: ArgArray, cons_equal: FuncArray, cons_unequal: FuncArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="gradient_descent", lamk: float=6, muk: float=10, sigma: float=8, alpha: float=0.5, beta: float=0.7, p: float=2, eta: float=1e-3, epsilon: float=1e-4, k: int=0) -> OutputType | 增广拉格朗日乘子法 |


```python
Expand All @@ -114,9 +114,3 @@ oc.mixequal.penalty_L1(f, (x1, x2), c1, c2, (1.5, 0.5))

(array([2., 1.]), 47)




```python

```
10 changes: 5 additions & 5 deletions examples/doc/_example.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -20,10 +20,10 @@ oe.Lasso.[函数名]([矩阵A], [矩阵b], [因子mu], [参数表], [初始迭

| 方法头 | 解释 |
| ------------------------------------------------------------------------------------------------------- | ---------------- |
| gradient(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, delta: Optional[float]=10, alp: Optional[float]=1e-3, epsilon: Optional[float]=1e-2, k: Optional[int]=0) -> OutputType | 光滑化Lasso函数法 |
| subgradient(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, alphak: Optional[float]=2e-2, epsilon: Optional[float]=1e-3, k: Optional[int]=0) -> OutputType | 次梯度法Lasso避免一阶不可导 |
| penalty(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, gamma: Optional[float]=0.01, epsilon: Optional[float]=1e-6, k: Optional[int]=0) -> OutputType | 罚函数法 |
| approximate_point(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, epsilon: Optional[float]=1e-4, k: Optional[int]=0) -> OutputType | 邻近算子更新 |
| gradient(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, delta: float=10, alp: float=1e-3, epsilon: float=1e-2, k: int=0) -> OutputType | 光滑化Lasso函数法 |
| subgradient(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, alphak: float=2e-2, epsilon: float=1e-3, k: int=0) -> OutputType | 次梯度法Lasso避免一阶不可导 |
| penalty(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, gamma: float=0.01, epsilon: float=1e-6, k: int=0) -> OutputType | 罚函数法 |
| approximate_point(A: NDArray, b: NDArray, mu: float, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, epsilon: float=1e-4, k: int=0) -> OutputType | 邻近算子更新 |


```python
Expand Down Expand Up @@ -69,7 +69,7 @@ oe.WanYuan.[函数名]([直线的斜率], [直线的截距], [二次项系数],

| 方法头 | 解释 |
| --------------------------------------------------------------- | -------------------- |
| solution(m: float, n: float, a: float, b: float, c: float, x3: float, y3: float, x_0: tuple, draw: Optional[bool]=False, eps: Optional[float]=1e-10) -> None | 使用高斯-牛顿方法求解构造的7个残差函数 |
| solution(m: float, n: float, a: float, b: float, c: float, x3: float, y3: float, x_0: tuple, draw: bool=False, eps: float=1e-10) -> None | 使用高斯-牛顿方法求解构造的7个残差函数 |


```python
Expand Down
24 changes: 12 additions & 12 deletions examples/doc/_unconstrain.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -33,9 +33,9 @@ ou.gradient_descent.[函数名]([目标函数], [参数表], [初始迭代点])

| 方法头 | 解释 |
| ----------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------ |
| solve(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 通过解方程的方式来求解精确步长 |
| steepest(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 使用线搜索方法求解非精确步长（默认使用wolfe线搜索） |
| barzilar_borwein(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="Grippo", c1: Optional[float]=0.6, beta: Optional[float]=0.6, alpha: Optional[float]=1, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 使用Grippo与ZhangHanger提出的非单调线搜索方法更新步长 |
| solve(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, epsilon: float=1e-10, k: int=0) -> OutputType | 通过解方程的方式来求解精确步长 |
| steepest(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", epsilon: float=1e-10, k: int=0) -> OutputType | 使用线搜索方法求解非精确步长（默认使用wolfe线搜索） |
| barzilar_borwein(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="Grippo", c1: float=0.6, beta: float=0.6, alpha: float=1, epsilon: float=1e-10, k: int=0) -> OutputType | 使用Grippo与ZhangHanger提出的非单调线搜索方法更新步长 |


```python
Expand Down Expand Up @@ -65,9 +65,9 @@ ou.newton.[函数名]([目标函数], [参数表], [初始迭代点])

| 方法头 | 解释 |
| ----------------------------------------------------------------------------------------------- | --------------------------------- |
| classic(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 通过直接对目标函数二阶导矩阵（海瑟矩阵）进行求逆来获取下一步的步长 |
| modified(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", m: Optional[int]=20, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 修正当前海瑟矩阵保证其正定性（目前只接入了一种修正方法） |
| CG(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", epsilon: Optional[float]=1e-6, k: Optional[int]=0) -> OutputType | 采用牛顿-共轭梯度法求解梯度（非精确牛顿法的一种） |
| classic(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, epsilon: float=1e-10, k: int=0) -> OutputType | 通过直接对目标函数二阶导矩阵（海瑟矩阵）进行求逆来获取下一步的步长 |
| modified(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", m: int=20, epsilon: float=1e-10, k: int=0) -> OutputType | 修正当前海瑟矩阵保证其正定性（目前只接入了一种修正方法） |
| CG(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", epsilon: float=1e-6, k: int=0) -> OutputType | 采用牛顿-共轭梯度法求解梯度（非精确牛顿法的一种） |


```python
Expand Down Expand Up @@ -97,9 +97,9 @@ ou.newton_quasi.[函数名]([目标函数], [参数表], [初始迭代点])

| 方法头 | 解释 |
| -------------------------------------------------------------------------------------------- | --------------- |
| bfgs(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", m: Optional[float]=20, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | BFGS方法更新海瑟矩阵 |
| dfp(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", m: Optional[float]=20, epsilon: Optional[float]=1e-4, k: Optional[int]=0) -> OutputType | DFP方法更新海瑟矩阵 |
| L_BFGS(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", m: Optional[float]=6, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 双循环方法更新BFGS海瑟矩阵 |
| bfgs(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", m: float=20, epsilon: float=1e-10, k: int=0) -> OutputType | BFGS方法更新海瑟矩阵 |
| dfp(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", m: float=20, epsilon: float=1e-4, k: int=0) -> OutputType | DFP方法更新海瑟矩阵 |
| L_BFGS(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", m: float=6, epsilon: float=1e-10, k: int=0) -> OutputType | 双循环方法更新BFGS海瑟矩阵 |



Expand Down Expand Up @@ -130,8 +130,8 @@ ou.nonlinear_least_square.[函数名]([目标函数], [参数表], [初始迭代

| 方法头 | 解释 |
| ---------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------- |
| gauss_newton(funcr: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, method: Optional[str]="wolfe", epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | 高斯-牛顿提出的方法框架，包括OR分解等操作 |
| levenberg_marquardt(funcr: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, m: Optional[float]=100, lamk: Optional[float]=1, eta: Optional[float]=0.2, p1: Optional[float]=0.4, p2: Optional[float]=0.9, gamma1: Optional[float]=0.7, gamma2: Optional[float]=1.3, epsilon: Optional[float]=1e-10, k: Optional[int]=0) -> OutputType | Levenberg Marquardt提出的方法框架 |
| gauss_newton(funcr: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, method: str="wolfe", epsilon: float=1e-10, k: int=0) -> OutputType | 高斯-牛顿提出的方法框架，包括OR分解等操作 |
| levenberg_marquardt(funcr: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, m: float=100, lamk: float=1, eta: float=0.2, p1: float=0.4, p2: float=0.9, gamma1: float=0.7, gamma2: float=1.3, epsilon: float=1e-10, k: int=0) -> OutputType | Levenberg Marquardt提出的方法框架 |


```python
Expand Down Expand Up @@ -164,7 +164,7 @@ ou.trust_region.[函数名]([目标函数], [参数表], [初始迭代点])

| 方法头 | 解释 |
| ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------- |
| steihaug_CG(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: Optional[bool]=True, output_f: Optional[bool]=False, m: Optional[float]=100, r0: Optional[float]=1, rmax: Optional[float]=2, eta: Optional[float]=0.2, p1: Optional[float]=0.4, p2: Optional[float]=0.6, gamma1: Optional[float]=0.5, gamma2: Optional[float]=1.5, epsilon: Optional[float]=1e-6, k: Optional[int]=0) -> OutputType | 截断共轭梯度法在此方法中被用于搜索步长 |
| steihaug_CG(funcs: FuncArray, args: ArgArray, x_0: PointArray, draw: bool=True, output_f: bool=False, m: float=100, r0: float=1, rmax: float=2, eta: float=0.2, p1: float=0.4, p2: float=0.6, gamma1: float=0.5, gamma2: float=1.5, epsilon: float=1e-6, k: int=0) -> OutputType | 截断共轭梯度法在此方法中被用于搜索步长 |


```python
Expand Down
2 changes: 1 addition & 1 deletion optimtool/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -28,4 +28,4 @@
from ._version import __version__

if sys.version_info < (3, 7, 0):
raise OSError(f'optimtool-2.4.0 requires Python >=3.7, but yours is {sys.version}')
raise OSError(f'optimtool-2.4.1 requires Python >=3.7, but yours is {sys.version}')
Loading

0 comments on commit c78f7f7

Please sign in to comment.