文章快速检索 高级检索
0

页岩气专题论文

引用本文 [复制中英文]

张毅. 相空间中非保守系统的Herglotz广义变分原理及其Noether定理1)[J]. 力学学报, 2016, 48(6): 1382-1389. DOI: 10.6052/0459-1879-16-086.
[复制中文]
Zhang Yi. GENERALIZED VARIATIONAL PRINCIPLE OF HERGLOTZ TYPE FOR NONCONSERVATIVE SYSTEM IN PHASE SPACE AND NOETHER'S THEOREM1)[J]. Chinese Journal of Theoretical and Applied Mechanics, 2016, 48(6): 1382-1389. DOI: 10.6052/0459-1879-16-086.
[复制英文]

基金项目

1) 国家自然科学基金资助项目(11272227,11572212)

通讯作者

2) 张毅, 教授, 主要研究方向:分析力学.E-mail:zhy@mail.usts.edu.cn

文章历史

2016-04-05 收稿
2016-08-04 录用
2016-08-15网络版发表
相空间中非保守系统的Herglotz广义变分原理及其Noether定理1)
张毅2)     
苏州科技大学土木工程学院, 苏州 215011
摘要: 与经典变分原理相比,基于由微分方程定义的作用量的Herglotz广义变分原理给出了非保守动力学系统的一个变分描述,它不仅能够描述所有采用经典变分原理能够描述的动力学过程,而且能够应用于经典变分原理不能适用的非保守或耗散系统.将Herglotz广义变分原理拓展到相空间,研究相空间中非保守力学系统的Herglotz广义变分原理与Noether定理及其逆定理.首先,提出相空间中Herglotz广义变分原理,给出相空间中非保守系统的变分描述,导出相应的Hamilton正则方程;其次,基于非等时变分与等时变分之间的关系,导出相空间中Hamilton-Herglotz作用量变分的两个基本公式;再次,给出Noether对称变换的定义和判据,提出并证明相空间中非保守系统基于Herglotz变分问题的Noether定理及其逆定理,揭示了相空间中力学系统的Noether对称性与守恒量之间的内在联系.在经典条件下,Herglotz广义变分原理退化为经典变分原理,与之相应的相空间中的Noether定理退化为经典Hamilton系统的Noether定理.文末以著名的Emden方程和平方阻尼振子为例说明上述方法和结果的有效性.
关键词: Herglotz广义变分原理    Noether定理    非保守动力学    相空间    
引言

Noether定理首次从变分学背景下揭示了对称性与守恒量之间的相互关系[1],即Hamilton作用量在关于广义坐标和时间的变换群的无限小变换下的不变性意味着沿着系统的动力学真实运动轨道存在一个守恒量.Noether定理阐释了牛顿力学的所有守恒量,如:时间的均匀性导致质点系的能量守恒;空间的均匀性导致质点系的动量守恒;空间各向同性导致质点系的动量矩守恒[2].Djukić等[3]将结果拓展到完整非保守系统并考虑无限小变换的生成元包含广义速度的情形;李子平[4],Bahar等[5]给出线性非完整约束系统的Noether定理;刘端[6]进一步将结果拓展到一般非完整非保守力学系统;梅凤翔[7-8]建立了Birkhoff系统和广义Birkhoff系统的Noether理论;文献[9-10]研究了分数阶Birkhoff系统的Noether对称性与守恒量.关于Noether定理及其应用的研究已取得重要进展[11-23].但是,上述Noether定理都是基于经典形式的变分原理,即:变分问题的作用量是由积分泛函定义的,它们不能应用于作用量由微分方程定义的一般情形.

1930年,Herglotz[24]在研究接触变换及其与Hamilton系统和Poisson括号的联系时提出了一类广义变分原理,其作用量是由微分方程来定义的.Herglotz广义变分原理不仅能够描述所有采用经典变分原理能够描述的物理过程,而且能够应用于解决经典变分原理不能适用的问题,如:非保守动力学过程或耗散系统的变分描述[25-26].2002年,Georgieva等[27]研究了Herglotz广义变分原理和Noether定理.但是迄今为止基于Herglotz广义变分原理的Noether对称性的研究尚不多见且研究限于位形空间中的Lagrange系统[28-30].由于力学系统的相空间具有自然辛结构,在数学描述上比Lagrange力学要容易[31].对于某些系统,在位形空间中其对称性并不明显地表现出来,但在相空间中却具有一定的对称性质,用正则形式的Noether定理则可导出相应的守恒量[2].本文将研究相空间中非保守系统的Herglotz广义变分原理,并基于非等时变分与等时变分之间的关系导出相空间中Hamilton-Herglotz作用量的变分公式,建立在此情形下Noether对称性的判据,提出并证明相空间中Herglotz变分问题的Noether定理及其逆定理.文末以著名的Emden方程和平方阻尼振子系统为例说明结果的应用.

1 相空间中Herglotz广义变分原理与Hamilton正则方程

根据位形空间中Herglotz广义变分原理[29],相空间中非保守系统的Herglotz变分问题可定义为:

确定函数$q_s (t)$$p_s (t)$,使由微分方程

$ \dot {z}(t) = p_s (t)\dot {q}_s (t) - H\left( {t, q_s (t), p_s (t), z(t)} \right) $ (1)

定义的泛函$z(t)$,在给定的端点条件

$ \left. {q_s (t)} \right|_{t = t_1 } = q_{s1} \, , \ \ \left. {q_s \left( t \right)} \right|_{t = t_2 } = q_{s2} \, \ \ \left( {s = 1, 2, \cdots, n} \right) $ (2)

和初始条件

$ \left. {z(t)} \right|_{t = t_1 } = z_1 $ (3)

下,在$t=t_2$时取得极值.其中$q_s \left( t \right)$$p_s (t) \, \left( {s = 1, 2, \cdots, n} \right)$分别为系统的广义坐标和广义动量,$H\left( {t, q_k, p_k, z} \right)$可称为Hamilton函数,$q_{s1}$$q_{s2}$$z_1 $均为固定常数.

由式(1)确定的泛函z可称为Hamilton-Herglotz作用量,上述变分问题可称为相空间中非保守系统的Herglotz广义变分原理.

对方程(1)进行等时变分运算,有

$ \delta \dot {z} = \dot {q}_s \delta p_s + p_s \delta \dot {q}_s - \dfrac{\partial H}{\partial q_s }\delta q_s -\\ \qquad \dfrac{\partial H}{\partial p_s }\delta p_s - \dfrac{\partial H}{\partial z}\delta z $ (4)

这里及文中采用Einstein求和约定,即同一项中两个相同的活动指标表示对其求和.利用交换关系$\frac{{{\rm{d}}\delta z}}{{{\rm{d}}t}} = \delta \dot {z}$,式(4)可写为

$ \frac{{{\rm{d}}\delta z}}{{{\rm{d}}t}} = A - \dfrac{\partial H}{\partial z}\delta z $ (5)

其中

$ A = \dot {q}_s \delta p_s + p_s \delta \dot {q}_s - \dfrac{\partial H}{\partial q_s }\delta q_s - \dfrac{\partial H}{\partial p_s }\delta p_s $ (6)

方程(5)是关于$\delta z$的一阶线性非齐次常微分方程,其解可表示为

$ \delta z(t)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) - \delta z\left( {t_1 } \right) = \\ \qquad \int_{t_1 }^t {A\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}} t } } \right){\rm{d}}t} $ (7)

由初始条件(3),且考虑到$z(t)$$t = t_2 $取得极值,有

$ \delta z\left( {t_1 } \right) = \delta z\left( {t_2 } \right) = 0 $ (8)

方程(7)在所有$t \in \left[{t_1, t_2 } \right]$成立,特别地,取$t = t_2$,有

$ \int_{t_1 }^{t_2 } {A\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right){\rm{d}}t} = 0 $ (9)

将式(6)代入方程(9),对$\delta \dot {q}_s $进行分部积分运算,并利用方程(1)和端点条件(2),得到

$ \begin{array}{l} \int_{{t_1}}^{{t_2}} {\exp \left( {\int_{{t_1}}^t {\frac{{\partial H}}{{\partial z}}{\rm{d}}t} } \right)} \left[ {\left( { - {{\dot p}_s} - \frac{{\partial H}}{{\partial {q_s}}} - {p_s}\frac{{\partial H}}{{\partial z}}} \right)\delta {q_s} + } \right.\\ \;\;\;\;\;\;\;\left. {\left( {{{\dot q}_s} - \frac{{\partial H}}{{\partial {p_s}}}} \right)\delta {p_s}} \right]{\rm{d}}t = 0 \end{array} $ (10)

$\delta q_s$$\delta p_s $的独立性,并利用变分学基本引理[32],由方程(10)得到

$ \left. \begin{array}{l} \exp \left( {\int_{{t_1}}^t {\frac{{\partial H}}{{\partial z}}{\rm{d}}t} } \right)\left( {{{\dot q}_s} - \frac{{\partial H}}{{\partial {p_s}}}} \right) = 0\\ \exp \left( {\int_{{t_1}}^t {\frac{{\partial H}}{{\partial z}}{\rm{d}}t} } \right)\left( { - {{\dot p}_s} - \frac{{\partial H}}{{\partial {q_s}}} - {p_s}\frac{{\partial H}}{{\partial z}}} \right) = 0\\ \;\;\;\;\;\;\left( {s = 1,2, \cdots ,n} \right) \end{array} \right\} $ (11)

$ \dot {q}_s = \dfrac{\partial H}{\partial p_s }\, , \ \ \dot {p}_s = - \dfrac{\partial H}{\partial q_s } - p_s \dfrac{\partial H}{\partial z} \ \ \left( {s = 1, 2, \cdots, n} \right) $ (12)

方程(12)可称为相空间中非保守系统基于Herglotz广义变分原理的Hamilton正则方程.

如果Hamilton函数不显含z,即$H = H\left( {t, q_k, p_k } \right)$,则方程(1)成为

$ z = \int_{t_1 }^{t_2 } {\left[{p_s (t)\dot {q}_s \left( t \right)-H\left( {t, q_s (t), p_s (t)} \right)} \right]{\rm{d}}t} $ (13)

而方程(12)成为

$ \dot {q}_s = \dfrac{\partial H}{\partial p_s } \, , \ \ \dot {p}_s = - \dfrac{\partial H}{\partial q_s } \ \ \ \left( {s = 1, 2, \cdots, n} \right) $ (14)

式(13)是相空间中Hamilton作用量,方程(14)为经典Hamilton正则方程.

方程(12)中第二组方程的项$\left( {p_s \dfrac{\partial H}{\partial z}} \right)$对应系统的非保守力.与以往的非保守系统的Hamilton原理[33]不同,相空间中Herglotz广义变分原理是极值原理,它提供了相空间中非保守力学系统的一个变分描述.利用Herglotz广义变分原理及其Hamilton正则方程(12)可以系统地处理保守和非保守问题,并且经典Hamilton原理和经典Hamilton正则方程是其特例.

2 Hamilton-Herglotz作用量变分的两个基本公式

引进时间t,广义坐标$q_s$和广义动量$p_s$r参数有限变换群$G_r$的无限小变换

$ \left. \begin{array}{l} \bar t = t + \Delta t\\ {{\bar q}_s}\left( {\bar t{\mkern 1mu} } \right) = {q_s}(t) + \Delta {q_s}\;\;\left( {s = 1,2, \cdots ,n} \right)\\ {{\bar p}_s}\left( {\bar t{\mkern 1mu} } \right) = {p_s}(t) + \Delta {p_s} \end{array} \right\} $ (15)

或其展开式

$ \left. \begin{array}{l} \bar t = t + {\varepsilon _\sigma }{\tau ^\sigma }\left( {t,{q_k},{p_k},z} \right)\\ {{\bar q}_s}\left( {\bar t{\mkern 1mu} } \right) = {q_s}(t) + {\varepsilon _\sigma }\xi _s^\sigma \left( {t,{q_k},{p_k},z} \right)\\ {{\bar p}_s}\left( {\bar t{\mkern 1mu} } \right) = {p_s}(t) + {\varepsilon _\sigma }\eta _s^\sigma \left( {t,{q_k},{p_k},z} \right)\\ \;\;\;\;\;\;\left( {s,k = 1,2, \cdots ,n} \right) \end{array} \right\} $ (16)

其中$\Delta \left( * \right)$表示$\left( * \right)$的非等时变分,$\varepsilon _\sigma \left( {\sigma = 1, 2, \cdots, r} \right)$为无限小参数,$\tau ^\sigma$$\xi _s^\sigma $$\eta _s^\sigma $为无限小变换的生成函数或生成元.在变换(15)作用下,Hamilton-Herglotz作用量z变为

$ \bar {z}\left( \bar {t} \, \right) = z(t) + \Delta z\left( t \right) $ (17)

作用量z的非等时变分$\Delta z$为变换前后作用量z之差相对$\varepsilon _\sigma $的主线性部分.对于任意函数F,非等时变分$\Delta F$与等时变分$\delta F$之间存在关系[2]

$ \Delta F = \delta F + \dot {F}\Delta t $ (18)

由式(18),并注意到交换关系

$ \frac{{{\rm{d}}\delta F}}{{{\rm{d}}t}} = \delta \frac{{{\rm{d}}F}}{{{\rm{d}}t}} = \delta \dot F $ (19)

易得

$ \Delta \dot F = \frac{{{\rm{d}}\Delta F}}{{{\rm{d}}t}} - \dot F\frac{{{\rm{d}}\Delta t}}{{{\rm{d}}t}} $ (20)

由方程(1),得到

$ \Delta \dot {z} = \dot {q}_s \Delta p_s + p_s \Delta \dot {q}_s - \dfrac{\partial H}{\partial t}\Delta t - \\ \qquad \dfrac{\partial H}{\partial q_s }\Delta q_s - \dfrac{\partial H}{\partial p_s }\Delta p_s - \dfrac{\partial H}{\partial z}\Delta z $ (21)

利用式(20),并考虑到方程(1),式(21)成为

$ \dfrac{{\rm{d}} \Delta z }{{\rm{d}}t} = \dot {q}_s \Delta p_s + p_s \dfrac{{\rm{d}} \Delta q_s }{{\rm{d}}t} - H\dfrac{{\rm{d}} \Delta t }{{\rm{d}}t} - \dfrac{\partial H}{\partial t}\Delta t - \\ \qquad \dfrac{\partial H}{\partial q_s }\Delta q_s - \dfrac{\partial H}{\partial p_s }\Delta p_s - \dfrac{\partial H}{\partial z}\Delta z $ (22)

解方程(22),可得

$ \Delta z(t)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) - \Delta z\left( {t_1 } \right) = \\ \qquad \int_{t_1 }^t {\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)} \left( {\dot {q}_s \Delta p_s + p_s \dfrac{{\rm{d}} \Delta q_s }{{\rm{d}}t} - H\dfrac{{\rm{d}} \Delta t }{{\rm{d}}t} - }\right.\\ \qquad \left.{ \dfrac{\partial H}{\partial t}\Delta t - \dfrac{\partial H}{\partial q_s }\Delta q_s - \dfrac{\partial H}{\partial p_s }\Delta p_s } \right){\rm{d}}t $ (23)

显然$\Delta z\left( {t_1 } \right) = 0$,方程(23)可表示为

$ \Delta z(t)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) = \\ \quad \int_{t_1 }^t \left\{ {\dfrac{{\rm{d}} }{{\rm{d}}t}\left[{\left( {p_s \Delta q_s-H\Delta t} \right)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)} \right]} \right. +\\ \quad \exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)\left[{\left( {- \dot {p}_s- \dfrac{\partial H}{\partial q_s }- p_s \dfrac{\partial H}{\partial z}} \right)\left( {\Delta q_s - \dot {q}_s \Delta t} \right) + }\right.\\ \quad \left.{\left.{\left( {\dot {q}_s -\dfrac{\partial H}{\partial p_s }} \right)\left( {\Delta p_s -\dot {p}_s \Delta t} \right)} \right]} \right\} {\rm{d}}t $ (24)

由于

$ \Delta t = \varepsilon _\sigma \tau ^\sigma\, , \ \ \Delta q_s = \varepsilon _\sigma \xi _s^\sigma \\ \Delta p_s = \varepsilon _\sigma \eta _s^\sigma \quad \left( {s = 1, 2, \cdots, n} \right) $ (25)

将式(25)代入式(23)和式(24),得到

$ \Delta z(t)\exp \left( {\int_{t_1 }^t \dfrac{\partial H}{\partial z}{\rm{d}}t } \right) =\\ \qquad \int_{t_1 }^t \left[{\exp \left( {\int_{t_1 }^t \dfrac{\partial H}{\partial z}{\rm{d}}t } \right)\left( {\dot {q}_s \eta _s^\sigma + p_s \dot {\xi }_s^\sigma- H\dot {\tau }^\sigma- }\right.}\right.\\ \qquad \left.{\left.{\dfrac{\partial H}{\partial t}\tau ^\sigma - \dfrac{\partial H}{\partial q_s }\xi _s^\sigma -\dfrac{\partial H}{\partial p_s }\eta _s^\sigma } \right)} \right] \varepsilon _\sigma {\rm{d}}t $ (26)

以及

$ \Delta z(t)\exp \left( {\int_{t_1 }^t \dfrac{\partial H}{\partial z}{\rm{d}}t } \right)= \\ \quad \int_{t_1 }^t \left\{ {\dfrac{{\rm{d}} }{{\rm{d}}t}\left[{\left( {p_s \xi _s^\sigma-H\tau ^\sigma } \right)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)} \right]} \right. +\\ \quad \left.{ \exp \left( {\int_{t_1 }^t \dfrac{\partial H}{\partial z}{\rm{d}}t } \right)\left[{\left( {- \dot {p}_s- \dfrac{\partial H}{\partial q_s }- p_s \dfrac{\partial H}{\partial z}} \right)\left( {\xi _s^\sigma - \dot {q}_s \tau ^\sigma } \right) +}\right.}\right. \\ \quad \left.{\left.{ \left( {\dot {q}_s -\dfrac{\partial H}{\partial p_s }} \right)\left( {\eta _s^\sigma - \dot {p}_s \tau ^\sigma } \right)} \right]} \right\} \varepsilon _\sigma {\rm{d}}t $ (27)

式(26)和式(27)是相空间中Hamilton-Herglotz作用量变分的两个基本公式.

3 相空间中基于Herglotz变分问题的Noether定理

下面根据Noether对称性的概念[12],给出相空间中非保守系统基于Herglotz变分问题的Noether对称变换的定义和判据.

定义1    对于相空间中非保守系统的Herglotz变分问题,如果Hamilton -Herglotz作用量是无限小群变换的不变量,即:对每一个无限小变换,始终成立

$ \Delta z\left( {t_2 } \right) = 0 $ (28)

则称无限小群变换为系统(12)的Noether对称变换.

由定义1和式(26),可以得到如下判据:

判据1    对于无限小群变换(16),如果无限小生成元$\tau ^\sigma $$\xi _s^\sigma $$\eta _s^\sigma $满足条件

$ \dot {q}_s \eta _s^\sigma + p_s \dot {\xi }_s^\sigma - H\dot {\tau }^\sigma -\dfrac{\partial H}{\partial t}\tau ^\sigma - \\ \qquad \dfrac{\partial H}{\partial q_s }\xi _s^\sigma - \dfrac{\partial H}{\partial p_s }\eta _s^\sigma = 0 \quad \left( {\sigma = 1, 2, \cdots, r} \right) $ (29)

则变换为相空间中非保守系统基于Herglotz变分问题的Noether对称变换.

下面建立相空间中非保守系统基于Herglotz变分问题的Noether定理,有:

定理1    对于相空间中非保守系统的Herglotz变分问题,如果无限小群变换(16)是系统(12)的Noether对称变换,则该系统存在r个线性独立的守恒量,形如

$ \left\{ \begin{array}{l} I_N^\sigma = \left( {{p_s}\xi _s^\sigma - H{\tau ^\sigma }} \right)\exp \left( {\int_{{t_1}}^t {\frac{{\partial H}}{{\partial z}}{\rm{d}}t} } \right) = {c^\sigma }\\ \left( {\sigma = 1,2, \cdots ,r} \right) \end{array} \right. $ (30)

证明    因为无限小群变换(16)是系统(12)的Noether对称变换,由定义1,有

$ \Delta z\left( {t_2 } \right) = 0 $

将上式代入式(27),得

$ \int_{t_1 }^{t_2 } {\left\{ {\dfrac{{\rm{d}} }{{\rm{d}}t}\left[{\left( {p_s \xi _s^\sigma-H\tau ^\sigma } \right)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)} \right]} \right.} + \\ \quad \exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) \cdot \left[{\left( {- \dot {p}_s- \dfrac{\partial H}{\partial q_s }- p_s \dfrac{\partial H}{\partial z}} \right)\left( {\xi _s^\sigma - \dot {q}_s \tau ^\sigma } \right) +}\right.\\ \quad \left.{ \left. { \left( {\dot {q}_s -\dfrac{\partial H}{\partial p_s }} \right)\left( {\eta _s^\sigma -\dot {p}_s \tau ^\sigma } \right)} \right]} \right\} \varepsilon _\sigma {\rm{d}}t = 0 $ (31)

将方程(12)代入式(31),并考虑到$\varepsilon _\sigma $的独立性和积分区间的任意性,得

$ \left. \begin{array}{l} \frac{{\rm{d}}}{{{\rm{d}}t}}\left[ {\left( {{p_s}\xi _s^\sigma - H{\tau ^\sigma }} \right)\exp \left( {\int_{{t_1}}^t {\frac{{\partial H}}{{\partial z}}{\rm{d}}t} } \right)} \right] = 0\\ \left( {\sigma = 1,2, \cdots ,r} \right) \end{array} \right\} $ (32)

积分之,便得式(30).证毕.

定理1可称为相空间中非保守系统基于Herglotz变分问题的Noether定理.利用该定理可以通过Noether对称性找到非保守动力学系统或耗散系统的守恒量.如果Hamilton函数不显含z,则定理1成为

推论1    对于经典Hamilton系统(14),如果无限小群变换是系统的Noether对称变换,则系统存在r个线性独立的守恒量,形如

$ I_N^\sigma = p_s \xi _s^\sigma - H\tau ^\sigma = c^\sigma \quad \left( {\sigma = 1, 2, \cdots, r} \right) $ (33)

推论1是经典Hamilton系统的Noether定理[12].

4 相空间中基于Herglotz变分问题的Noether逆定理

假设系统(12)有r个线性独立的守恒量

$ I^\sigma = I^\sigma \left( {t, q_s, p_s } \right) = \mbox{const} \ \ \ \left( {\sigma = 1, 2, \cdots, r} \right) $ (34)

将式(34)对时间t求导,得到

$ \dfrac{{\rm{d}} I^\sigma }{{\rm{d}}t} = \dfrac{\partial I^\sigma }{\partial t} + \dfrac{\partial I^\sigma }{\partial q_s }\dot {q}_s + \dfrac{\partial I^\sigma }{\partial p_s }\dot {p}_s = 0 $ (35)

将Hamilton正则方程(11)的第二组方程乘以$\left( {\xi _s^\sigma - \dot {q}_s \tau ^\sigma } \right)$并对s求和,得

$ \exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)\left( { - \dot {p}_s - \dfrac{\partial H}{\partial q_s } - p_s \dfrac{\partial H}{\partial z}} \right)\left( {\xi _s^\sigma - \dfrac{\partial H}{\partial p_s }\tau ^\sigma } \right) = 0 $ (36)

比较式(35)和式(36)中含$\dot {p}_s $的项,得到

$ \exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right)\left( {\xi _s^\sigma - \dfrac{\partial H}{\partial p_s }\tau ^\sigma } \right) = \dfrac{\partial I^\sigma }{\partial p_s } $

$ \xi _s^\sigma = \dfrac{\partial H}{\partial p_s }\tau ^\sigma + \dfrac{\partial I^\sigma }{\partial p_s }\exp \left( { - \int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) \\ \qquad \left( {s = 1, 2, \cdots, n\, ; \ \sigma = 1, 2, \cdots, r} \right) $ (37)

再令已知积分(34)等于守恒量(30),即

$ \left( {p_s \xi _s^\sigma - H\tau ^\sigma } \right)\exp \left( {\int_{t_1 }^t {\dfrac{\partial H}{\partial z}{\rm{d}}t } } \right) = I^\sigma \\ \qquad \left( {\sigma = 1, 2, \cdots, r} \right) $ (38)

方程(37)和(38)是关于$\left( {n + 1} \right)r$个变量$\tau ^\sigma $$\xi _s^\sigma $$\left( {n + 1} \right)r$个代数方程,由此可找到生成元$\tau ^\sigma $$\xi _s^\sigma $,它们相应于系统的Noether对称变换.于是有

定理2    如果已知系统(12)的r个线性独立的守恒量(34),则由式(37)和式(38)确定的无限小群变换是系统的Noether对称变换.

定理2可称为相空间中Herglotz变分问题的Noether逆定理.利用该定理可由已知积分找到相应的Noether对称性.

如果Hamilton函数不显含z,则式(37)和式(38)成为

$ \xi _s^\sigma = \dfrac{\partial H}{\partial p_s }\tau ^\sigma + \dfrac{\partial I^\sigma }{\partial p_s }\, \ \ \left( {s = 1, 2, \cdots, n\, ; \ \sigma = 1, 2, \cdots, r} \right) $ (39)
$ p_s \xi _s^\sigma - H\tau ^\sigma = I^\sigma \, \ \ \left( {\sigma = 1, 2, \cdots, r} \right) $ (40)

于是定理2成为

推论2    如果已知经典Hamilton系统(14)的r个线性独立的守恒量(34),则由式(39)和式(40)确定的无限小群变换是系统的Noether对称变换.

推论2是经典Hamilton系统的Noether逆定理[12].

5 算例

例1    考虑著名的Emden方程[12]

$ \ddot {q} + \dfrac{2}{t}\dot {q} + q^5 = 0 $ (41)

它是一个单自由度完整非保守系统,试用本文方法研究其对称性与守恒量.

方程(41)可化为Herglotz变分问题来研究,其Lagrange函数为

$ L = \dfrac{1}{2}\dot {q}^2 - \dfrac{1}{6}q^6 - \dfrac{2}{t}z $ (42)

取广义动量和Hamilton函数为

$ p = \dfrac{\partial L}{\partial \dot {q}} = \dot {q}\, , \ \ H = p\dot {q} - L = \dfrac{1}{2}p^2 + \dfrac{1}{6}q^6 + \dfrac{2}{t}z $ (43)

其中作用量z满足微分方程

$ \dot {z} = p\dot {q} - \left( {\dfrac{1}{2}p^2 + \dfrac{1}{6}q^6 + \dfrac{2}{t}z} \right) $ (44)

方程(12)给出

$ \dot {q} = p\, , \ \ \dot {p} = - q^5 - p\dfrac{2}{t} $ (45)

方程(45)是方程(41)在Herglotz变分问题下的Hamilton正则形式.判据方程(29)给出

$ \dot {q}\eta + p\dot {\xi } - \left( {\dfrac{1}{2}p^2 + \dfrac{1}{6}q^6 + \dfrac{2}{t}z} \right)\dot {\tau } +\\ \qquad \dfrac{2}{t^2}z\tau - q^5\xi - p\eta = 0 $ (46)

方程(46)有解

$ \tau = - t - \dfrac{12zt}{3p^2t - q^6t - 12z} \\ \xi = \dfrac{1}{2}q - \dfrac{12pzt}{3p^2t - q^6t - 12z} \\ \eta = 1 $ (47)

生成元(47)相应于Emden方程(41)在相空间中基于Herglotz变分问题的Noether对称性,由定理1,得到

$ I_N = \dfrac{t^2} t_1^2 \left( {\dfrac{1}{2}pq + \dfrac{1}{2}tp^2 + \dfrac{1}{6}tq^6} \right) = \mbox{const}. $ (48)

式(48)是与生成元(47)相应的Noether守恒量.

其次,利用Noether逆定理由已知积分求相应的Noether对称变换.假设系统有积分(48),式(37)和式(38)分别给出

$ \xi = p\tau + \dfrac{1}{2}q + pt $ (49)
$ p\xi - \left( {\dfrac{1}{2}p^2 + \dfrac{1}{6}q^6 + \dfrac{2}{t}z} \right)\tau = \dfrac{1}{2}pq + \dfrac{1}{2}tp^2 + \dfrac{1}{6}tq^6 $ (50)

联立求解方程(49)和(50),得到

$ \tau = - t - \dfrac{12zt}{3p^2t - q^6t - 12z} \\ \xi = \dfrac{1}{2}q - \dfrac{12pzt}{3p^2t - q^6t - 12z} $ (51)

由定理2,生成元(51)相应于系统的Noether对称变换.

例2    研究平方阻尼振子,其运动微分方程为[12]

$ \ddot {q} + \gamma \dot {q}^2 + q = 0 $ (52)

其中$\gamma $为常数.此问题可化为Herglotz变分问题,其Lagrange函数为

$ L = \dfrac{1}{2}\dot {q}^2 - \dfrac{q}{2\gamma } + \dfrac{1}{4\gamma ^2} - 2\gamma \dot {q}z $ (53)

广义动量和Hamilton函数为

$ p = \dfrac{\partial L}{\partial \dot {q}} = \dot {q} - 2\gamma z \\ H = p\dot {q} - L = \dfrac{1}{2}\left( {p + 2\gamma z} \right)^2 + \dfrac{q}{2\gamma } -\dfrac{1}{4\gamma ^2} $ (54)

微分方程(1)给出

$ \dot {z} = \dfrac{1}{2}\left( {p + 2\gamma z} \right)^2 - 2\gamma z\left( {p + 2\gamma z} \right) - \dfrac{q}{2\gamma } + \dfrac{1}{4\gamma ^2} $ (55)

Hamilton正则方程(12)给出

$ \left. \begin{array}{l} \dot q = p + 2\gamma z\\ \dot p = - 2p\gamma \left( {p + 2\gamma z} \right) - \frac{1}{{2\gamma }} \end{array} \right\} $ (56)

判据方程(29)给出

$ \dot {q}\eta + p\dot {\xi } - \left[{\dfrac{1}{2}\left( {p + 2\gamma z} \right)^2 + \dfrac{q}{2\gamma }-\dfrac{1}{4\gamma ^2}} \right]\dot {\tau } - \\ \qquad \dfrac{1}{2\gamma }\xi - \left( {p + 2\gamma z} \right)\eta = 0 $ (57)

方程(57)有解

$ \tau = - 1\, , \ \ \xi = 0\, , \ \ \eta = t $ (58)

生成元(58)相应于平方阻尼振子系统(52)在相空间中基于Herglotz变分问题的Noether对称性,由定理1,得到

$ I_N = \dfrac{{\rm e}^{2\gamma q}}{{\rm e}^{2\gamma q_1 }}\left[{\dfrac{1}{2}\left( {p + 2\gamma z} \right)^2 + \dfrac{q}{2\gamma }-\dfrac{1}{4\gamma ^2}} \right] = \mbox{const} $ (59)

其中$q_1 \triangleq q\left( {t_1 } \right)$.式(59)是与生成元(58)相应的Noether守恒量.

其次,研究Noether逆定理的应用.假设系统有积分(59),由式(37)和式(38)得到

$ \left. \begin{array}{l} \xi = \left( {p + 2\gamma z} \right)\tau + p + 2\gamma z\\ p\xi - H\tau = \frac{1}{2}{\left( {p + 2\gamma z} \right)^2} + \frac{q}{{2\gamma }} - \frac{1}{{4{\gamma ^2}}} \end{array} \right\} $ (60)

由此解得

$ \tau = - 1\, , \ \ \ \xi = 0 $ (61)
6 结论

Herglotz广义变分原理为研究非保守或耗散系统动力学提供了一个有效途径,文章研究了相空间中非保守系统的Herglotz广义变分原理及其Noether对称性与守恒量.文章的主要贡献在于:提出了相空间中非保守系统的Herglotz广义变分原理,该原理给出了相空间中非保守系统的一个变分描述,建立了系统的Hamilton正则方程(12);导出了Hamilton-Herglotz作用量的非等时变分式(26)和式(27);建立了相空间中非保守系统基于Herglotz变分问题的Noether定理及其逆定理,即定理1和2.显然,通常的相空间中的Noether定理仅适用于作用量由积分泛函定义的情形,而不能用于作用量由微分方程定义的情形.但当Hamilton函数不显含z时,后者退化为前者.因此,定理包含了经典Hamilton系统的Noether定理和逆定理.文章的方法和结果可进一步推广到非完整系统,Birkhoff系统等.


参考文献
1 Noether AE. Invariante variationsprobleme. Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-Physikalische Klasse, 1918, KI Ⅱ:235-257
2 李子平. 经典和量子约束系统及其对称性质 . 北京: 北京工业大学出版社, 1993. ( Li Ziping. Classical and Quantal Dynamics of Constrained Systems and their Symmetrical Properties . Beijing: Beijing Polytechnic University Press, 1993. (in Chinese) )
3 Djukić Dj S, Vujanović B. Noether's theory in classical nonconservative mechanics. Acta Mechanica, 1975, 23 (1-2) : 17-27. DOI: 10.1007/BF01177666.
4 李子平. 约束系统的对称变换. 物理学报, 1981, 30 (12) : 1699-1706. ( Li Ziping. The transformation properties of constrained system. Acta Physica Sinica, 1981, 30 (12) : 1699-1706. (in Chinese) )
5 Bahar LY, Kwatny HG. Extension of Noether's theorem to constrained nonconservative dynamical systems. International Journal of Non-Linear Mechanics, 1987, 22 (2) : 125-138. DOI: 10.1016/0020-7462(87)90015-1.
6 Liu D. Noether's theorem and its inverse of nonholonomic nonconservative dynamical systems. Science in China (Series A), 1991, 34 (4) : 419-429.
7 Mei FX. The Noether's theory of Birkhoffan systems. Science in China (Serie A), 1993, 36 (12) : 1456-1467.
8 梅凤翔. 广义Birkhoff系统动力学 . 北京: 科学出版社, 2013. ( Mei Fengxiang. Dynamics of Generalized Birkhoffan System . Beijing: Science Press, 2013. (in Chinese) )
9 Zhang Y, Zhou Y. Symmetries and conserved quantities for fractional action-like Pfa an variational problems. Nonlinear Dynamics, 2013, 73 (1-2) : 783-793. DOI: 10.1007/s11071-013-0831-x.
10 Zhang Y, Zhai XH. Noether symmetries and conserved quantities for fractional Birkhoffan systems. Nonlinear Dynamics, 2015, 81 (1-2) : 469-480. DOI: 10.1007/s11071-015-2005-5.
11 梅凤翔. 分析力学(下卷) . 北京: 北京理工大学出版社, 2013. ( Mei Fengxiang. Analytical Mechanics (Ⅱ) . Beijing: Beijing Institute of Technology Press, 2013. (in Chinese) )
12 梅凤翔. 李群和李代数对约束力学系统的应用 . 北京: 科学出版社, 1999. ( Mei Fengxiang. Applications of Lie Groups and Lie Algebras to Constrained Mechanical Systems . Beijing: Science Press, 1999. (in Chinese) )
13 梅凤翔. 约束力学系统的对称性与守恒量 . 北京: 北京理工大学出版社, 2004. ( Mei Fengxiang. Symmetries and Conserved Quantities of Constrained Mechanical Systems . Beijing: Beijing Institute of Technology Press, 2004. (in Chinese) )
14 梅凤翔. 经典约束力学系统对称性与守恒量研究进展. 力学进展, 2009, 39 (1) : 37-43. ( Mei Fengxiang. Advances in the symmetries and conserved quantities of classical constrained systems. Advances in Mechanics, 2009, 39 (1) : 37-43. (in Chinese) )
15 Ferreira RAC, Malinowska AB. A counterexample to a Frederico-Torres fractional Noether-type theorem. Journal of Mathematical Analysis and Applications, 2015, 429 (2) : 1370-1373. DOI: 10.1016/j.jmaa.2015.03.060.
16 Atanacković TM, Konjik S, Pilipović S, et al. Variational problems with fractional derivatives:invariance conditions and Noether's theorem. Nonlinear Analysis, 2009, 71 (5-6) : 1504-1517. DOI: 10.1016/j.na.2008.12.043.
17 Malinowska AB, Torres DFM. Introduction to the Fractional Calculus of Variations . London: Imperial College Press, 2012.
18 Mei FX, Xie JF, Gang TQ. Weakly Noether symmetry for nonholonomic systems of Chetaev's type. Commun Theor Phys (Beijing, China), 2008, 49 (6) : 1413-1416. DOI: 10.1088/0253-6102/49/6/13.
19 Fu JL, Chen LQ, Chen BY. Noether-type theory for discrete mechanico-electrical dynamical systems with nonregular lattices. Science China:Physics, Mechanics, Astronomy, 2010, 53 (9) : 1687-1698. DOI: 10.1007/s11433-010-4043-9.
20 Long ZX, Zhang Y. Noether's theorem for fractional variational problem from El-Nabulsi extended exponentially fractional integral in phase space. Acta Mechanica, 2014, 225 (1) : 77-90. DOI: 10.1007/s00707-013-0956-5.
21 Zhai XH, Zhang Y. Noether symmetries and conserved quantities for fractional Birkhoffan systems with time delay. Communications in Nonlinear Science and Numerical Simulation, 2016, 36 : 81-97. DOI: 10.1016/j.cnsns.2015.11.020.
22 Song CJ, Zhang Y. Noether theorem for Birkhoffan systems on time scales. Journal of Mathematical Physics, 2015, 56 (10) : 102701. DOI: 10.1063/1.4932607.
23 Zhang Y, Zhou XS. Noether theorem and its inverse for nonlinear dynamical systems with nonstandard Lagrangians. Nonlinear Dynamics, 2016, 84 (4) : 1867-1876. DOI: 10.1007/s11071-016-2611-x.
24 Herglotz G. Berührungstransformationen. Lectures at the University of Göttingen, Göttingen, 1930
25 Georgieva B. Symmetries of the Herglotz variational principle in the case of one independent variable. Annual of Sofia University, the Faculty of Mathematics and Informatics, 2010, 100 : 113-122.
26 Santos SPS, Martins N, Torres DFM. Higher-order variational problems of Herglotz type. Vietnam Journal of Mathematics, 2014, 42 (4) : 409-419. DOI: 10.1007/s10013-013-0048-9.
27 Georgieva B, Guenther R. First Noether-type theorem for the generalized variational principle of Herglotz. Topological Methods in Nonlinear Analysis, 2002, 20 (2) : 261-273. DOI: 10.12775/TMNA.2002.036.
28 Georgieva B, Guenther R, Bodurov T. Generalized variational principle of Herglotz for several independent variables. First Noethertype theorem. Journal of Mathematical Physics, 2003, 44 (9) : 3911-3927. DOI: 10.1063/1.1597419.
29 Santos SPS, Martins N, Torres DFM. Variational problems of Herglotz type with time delay:Dubois-Reymond condition and Noether's first theorem. Discrete and Continuous Dynamical Systems, 2015, 35 (9) : 4593-4610. DOI: 10.3934/dcdsa.
30 Donchev V. Variational symmetries, conserved quantities and identities for several equations of mathematical physics. Journal of Mathematical Physics, 2014, 55 (3) : 032901. DOI: 10.1063/1.4867626.
31 梅凤翔. 经典力学从牛顿到伯克霍夫. 力学与实践, 1996, 18 (4) : 1-8. ( Mei Fengxiang. Classical mechanics from Newton to Birkhoffian. Mechanics in Engineering, 1996, 18 (4) : 1-8. (in Chinese) )
32 Goldstein H, Poole C, Safko J. Classical Mechanics (Third Edition) . Beijing: Higher Education Press, 2005.
33 梅凤翔, 刘端, 罗勇. 高等分析力学 . 北京: 北京理工大学出版社, 1991. ( Mei Fengxiang, Liu Duan, Luo Yong. Advanced Analytical Mechanics . Beijing: Beijing Institute of Technology Press, 1991. (in Chinese) )
GENERALIZED VARIATIONAL PRINCIPLE OF HERGLOTZ TYPE FOR NONCONSERVATIVE SYSTEM IN PHASE SPACE AND NOETHER'S THEOREM1)
Zhang Yi2)     
College of Civil Engineering, Suzhou University of Science and Technology, Suzhou 215011, China
Abstract: Compared with the classical variational principle, the generalized variational principle of Herglotz based upon the action defined by differential equations gives a variational description of nonconservative dynamical system. The principle can describe all dynamical processes and nonconservative or dissipative systems. In the present study, the principle is extended to phase space, and the generalized variational principle of Herglotz type for non-conservative mechanical system in phase space is given and Noether's theorem and its inverse of the system are studied. Firstly, the generalized variational principle of Herglotz type in phase space is presented, a variational description of non-conservative system in phase space is given, and the corresponding Hamilton canonical equations are deduced. Secondly, based upon the relation between non-isochronal variation and isochronal variation, two basic formulae for the variation of Hamilton-Herglotz action in phase space are obtained. Thirdly, the definition and the criterion of Noether symmetry are given, and Noether's theorem and its inverse of nonconservative system for the variational problem of Herglotz type in phase space are proposed and proved, and the inner relation between the Noether symmetry and the conserved quantity for mechanical systems in phase space is revealed. The generalized variational principle of Herglotz type reduces to the classical variational principle under classical conditions, and Noether's theorem for the variational problem of Herglotz type reduces to the classical Noether's theorem of Hamilton system. In the end of the paper, we take the famous Emden equation and damping oscillator with second power as examples to illustrate the application of the results.
Key words: generalized variational principle of Herglotz type    Noether's theorem    non-conservative dynamics    phase space