初值问题 dy/dx=y^(2/3)、y(0)=0 的解为什么不唯一？这涉及皮卡存在唯一性定理的哪个条件？
Why does the initial value problem dy/dx=y^(2/3), y(0)=0 fail to have a unique solution, and which hypothesis of the Picard existence-uniqueness theorem is involved?
常值函数 y≡0 是一个解；且对任意 c≥0，令 x≤c 时 y=0、x>c 时 y=(x-c)^3/27，两段光滑拼接且满足方程，故有无穷多解。根源是右端 f(y)=y^(2/3) 连续但对 y 的偏导 (2/3)y^(-1/3) 在 0 附近无界，破坏了局部利普希茨条件；皮卡-林德洛夫定理靠它收缩迭代锁定唯一解。
The constant function y≡0 is one solution, and for any c≥0 we may splice: set y=0 for x≤c and y=(x-c)^3/27 for x>c. The pieces join smoothly at c and satisfy the equation on each side, so there are infinitely many solutions. The culprit is f(y)=y^(2/3): it is continuous, yet its partial derivative in y, (2/3)y^(-1/3), blows up near y=0, so f fails to be locally Lipschitz in y. The Picard-Lindelof theorem relies on this condition to contract its iterates onto one solution; where it fails, uniqueness is lost.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
E. A. Coddington 与 N. Levinson，《Theory of Ordinary Differential Equations》（常微分方程理论），McGraw-Hill，1955，第1章第1-2节
E. A. Coddington & N. Levinson, Theory of Ordinary Differential Equations, McGraw-Hill, 1955, Chapter 1, Sections 1-2
独立核验依据 · Independent verification: 核验依据 1 ISBN 9780898747553 (1955)
#微分方程 #存在唯一性 #利普希茨条件
二阶线性微分方程在常点附近的解总能展开成普通幂级数；为什么遇到正则奇点就必须改用弗罗贝尼乌斯法 y = x^r Σ a_n x^n？它多捕捉到了什么？
Solutions of second-order linear ODEs near an ordinary point expand as ordinary power series; why must the Frobenius ansatz y = x^r Σ a_n x^n take over at a regular singular point, and what does it capture beyond power series?
常点处系数解析，递推顺利闭合；正则奇点处系数含极点，纯幂级数无法平衡最高阶导数的奇性。代入 y = x^r Σ a_n x^n 后最低次幂给出指标方程 r(r-1)+p₀r+q₀=0 定出 r，再逐阶求 a_n；两根差非整数得两个独立解，为整数或重根则第二解含 x^r ln x 型项，幂律与对数奇性一并捕获，贝塞尔方程即经典案例。
At an ordinary point the coefficients are analytic and recurrences close the power series neatly; at a regular singular point they carry poles, and a pure power series cannot balance the singularity of the highest derivative. Substituting y = x^r Σ a_n x^n, the lowest power yields the indicial equation r(r-1)+p₀r+q₀=0 fixing r, after which the a_n follow order by order. Roots differing by a non-integer give two independent solutions; an integer gap or double root forces a logarithmic term x^r ln x. The method captures power-law factors and logarithmic behavior alike, Bessel's equation being the classic case.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-04 · 2 条独立证据
来源 · Source
W. E. Boyce 与 R. C. DiPrima《Elementary Differential Equations and Boundary Value Problems》（第11版），Wiley，2017，第5章
W. E. Boyce & R. C. DiPrima, Elementary Differential Equations and Boundary Value Problems, 11th ed., Wiley, 2017, Chapter 5
独立核验依据 · Independent verification: 核验依据 1 ISBN 9780470115053 (2017) · 核验依据 2 ISBN 9780470115053 (2017)
#微分方程 #弗罗贝尼乌斯法 #正则奇点 #级数解
解一阶线性微分方程 y′ + p(x)y = q(x) 时，为什么两边乘以积分因子 μ(x) = exp(∫p(x)dx)，方程就变得可以直接积分？
When solving the first-order linear ODE y′ + p(x)y = q(x), why does multiplying through by the integrating factor μ(x) = exp(∫p(x)dx) turn it into something you can integrate right away?
关键在乘积求导法则：恒有 (μy)′ = μ′y + μy′。取 μ = exp(∫p dx) 使 μ′ = pμ，则 μ(y′ + py) 恰好合成单一导数 (μy)′，方程变为 (μy)′ = μq，积分得 μy = ∫μq dx + C。积分因子把耦合的两项打包成一个整体导数，故一阶线性方程总有显式通解公式。
The key is the product rule: (μy)′ = μ′y + μy′ holds identically. Choosing μ = exp(∫p dx) forces μ′ = pμ, so the left side μ(y′ + py) merges into the single derivative (μy)′; the equation becomes (μy)′ = μq, and integrating gives μy = ∫μq dx + C. An integrating factor packages two coupled unknown terms into one total derivative — which is precisely why every first-order linear equation admits an explicit general solution.
难度 进阶 · 适合 初中(12-15)
CC BY 4.0 · 已核验 · Verified 2026-10-04 · 2 条独立证据
来源 · Source
W. E. Boyce 与 R. C. DiPrima《Elementary Differential Equations and Boundary Value Problems》（第11版），Wiley，2017，§2.2
W. E. Boyce & R. C. DiPrima, Elementary Differential Equations and Boundary Value Problems, 11th ed., Wiley, 2017, §2.2
独立核验依据 · Independent verification: 核验依据 1 ISBN 9780470115053 (2017) · 核验依据 2 ISBN 9780470115053 (2017)
#微分方程 #积分因子 #一阶线性方程