---
title: Homotopy-Perturbation Method
url: https://www.emergentmind.com/topics/homotopy-perturbation-method-hpm
type: topic
---

# Homotopy-Perturbation Method

The Homotopy-Perturbation Method (HPM) is a semi-analytical technique for obtaining approximate or sometimes exact solutions of nonlinear ordinary and partial differential equations, including boundary-value, initial-value, and integro-differential problems. Combining homotopy theory from topology with the regular perturbation expansion, HPM provides a systematic procedure for reducing nonlinear problems to a sequence of linear subproblems, each typically solvable in closed form. This method, originally developed by J. H. He, has found widespread application across mathematical physics, engineering, astrophysics, and cosmology.

## 1. Structural Foundation and Methodology

HPM constructs a one-parameter family of deformations—termed a homotopy—between a simple linear problem and the original, typically nonlinear, target equation. Let the problem be written in operator form as
$$
A(u) = f,
$$
where $A$ is a (possibly nonlinear) differential operator. The operator is split as $A = L + N$ with $L$ linear and $N$ nonlinear. Introducing the homotopy (embedding) parameter $p\in[0,1]$, a continuous deformation is constructed:
$$
H(u,p) = (1-p)[L(u) - L(u_0)] + p[L(u) + N(u) - f] = 0,
$$
with $u_0$ an initial approximation, typically satisfying the boundary/initial conditions.

Assuming an analytic dependence on $p$, the solution is formally written as a power series,
$$
u(x,p) = u_0(x) + p\,u_1(x) + p^2\,u_2(x) + \ldots,
$$
and substituting into $H(u,p)=0$, one generates a recursive hierarchy of linear subproblems for the coefficients $u_k$. The approximate or exact solution is given by the sum at $p=1$,
$$
u(x) = \lim_{p\to1} u(x,p) = u_0(x) + u_1(x) + u_2(x) + \ldots
$$
This structural approach does not require the existence of a small physical parameter; rather, the formal $p$ organizes the analytic expansion around the linear reference point.

## 2. Theoretical Role within Homotopy-Based Methods

Recent theoretical advances demonstrate that HPM is a special, fixed-parameter case of the broader Homotopy Analysis Method (HAM), which provides additional convergence-control flexibility via choice of auxiliary linear operator and parameter(s) [2604.13063]. In HAM, the deformation equation includes tunable convergence-control ($c_0$) and auxiliary function ($H(\mathbf{r})$) degrees of freedom:
$$
(1-\varepsilon)\,\mathcal{L}[u(\varepsilon) - u_0] + \varepsilon\,c_0 H(\mathbf{r})\,\mathcal{F}(u(\varepsilon)) = 0\,,\quad \varepsilon \in [0,1].
$$
Fixing $\mathcal{L}=L$, $c_0=-1$, and $H=1$ recovers HPM exactly [2604.13063]. This embedding positions HPM as an analytically justified, but non-adaptive, degenerate case of HAM; while HPM provides simplicity and ease of use, it sacrifices tunable convergence control and robustness for strongly nonlinear systems.

## 3. Recursive Expansion and Implementation Details

The recursive structure of HPM is central to its implementation. After constructing the homotopy, substituting the power series into $H(u,p)$ and equating terms at each order in $p$ produces a hierarchy of linear problems:
* $O(p^0)$: typically yields $L(u_0) = L(u_0)$, vanishing identically or yielding $u_0$ as the initial approximation.
* $O(p^1)$: $L(u_1) + N(u_0) - f = 0$.
* $O(p^2)$: $L(u_2) + N'(u_0)u_1 = 0$ (with $N'$ the Fréchet derivative), and so forth.

Each subproblem is linear (in $u_k$), and higher-order corrections have homogeneous boundary conditions if $u_0$ is constructed to satisfy all (possibly multi-point) conditions [1310.2909]. The process is practically algorithmic, allowing direct algebraic or symbolic computation of each term.

For strongly nonlinear $N$, He’s polynomials or equivalent recursive expansions for products/powers are used [1804.05735], while for systems involving memory or fractional derivatives, auxiliary transforms (Laplace, Natural Transform) may be combined with HPM, yielding hybrid methods (e.g., HPTM, HNTHPM) [1611.06488, 1804.05735]. In all cases, the principle remains: nonlinearity appears only as explicit driving terms at each order, with convergence determined by the analyticity radius in $p$.

## 4. Convergence, Parameter Sensitivity, and Error Analysis

Convergence of the HPM series is guaranteed under analyticity assumptions and smallness of the nonlinear operator (in a Banach space), often implemented via the Implicit Function Theorem [2604.13063, 1310.2909]. For weakly nonlinear or analytic problems, the convergence radius in $p$ (i.e., the interval over which the $p$-series converges at $p=1$) is typically sufficient to ensure rapid convergence in practical computations. The lack of a convergence-control parameter in standard HPM, however, can result in divergence or slow convergence for strongly nonlinear or stiff problems, where an adaptive variant (e.g., HAM) is preferable [2604.13063].

In practical terms, convergence is monitored by the relative size of successive terms or by residual evaluation,
$$
\text{Res}[u] = L(u) + N(u) - f,
$$
computed for the truncated sum. Empirical demonstrations in domains such as population balances [1712.09013], cosmological models [1309.5596], and eigenvalue-boundary-value problems [1310.2909] show that only a few terms (3–6) are needed for high-precision approximation. Explicit error estimates for the truncated sum (up to order $\ell$) are possible when the operator norm admits a contractivity constant $\gamma<1$ [1611.06488].

## 5. Applications Across Mathematical Physics and Engineering

HPM has been successfully deployed in a vast range of nonlinear problems, including:
- **Astrophysics and General Relativity**: Analytically approximating neutron star structures via the TOV equation (yielding mass functions and compactness ranges) [1911.00350], calculation of gravitational waveforms [1203.5123], perihelion precession, and null-geodesic deflections (with exact or rapid-convergent series for orbits and deflection angles) [1612.07279, 1706.01809].
- **Cosmology**: Analytical Friedmann equation solutions, including dust and quintessence cosmologies [1309.5596], and luminosity distance-redshift relationships in FLRW models [1511.07459].
- **Fluid Mechanics and Applied Mathematics**: Blasius boundary layer, Burgers and KdV equations, MHD Jeffery-Hamel flow, heat transfer in nanofluids, where closed-form polynomial approximations for velocity, temperature, and other observables are constructed [2310.19525, 1601.05298, 1507.02790, 1904.13306].
- **Population Balances and Biomathematics**: Analytic series solutions for Smoluchowski-type aggregation and fragmentation equations [1712.09013].
- **Control Theory and Optimization**: Transformation of nonlinear optimal control problems into recursive linear two-point boundary-value problems, enabling explicit construction of state, costate, and optimal control law as convergent series [1409.4806].
- **Fractional PDEs**: Integration with Laplace or Natural Transform methods for time-fractional evolution and delay equations [1611.06488, 1804.05735].

In all such applications, HPM provides a framework where nonlinearity is handled recursively via driving terms, sidestepping the necessity for direct numerical iteration or small-parameter expansions.

## 6. Strengths, Limitations, and Methodological Extensions

**Strengths:**
- General applicability to both linear and nonlinear initial/boundary-value problems.
- Absence of any need for physical or artificial smallness parameter.
- Reduction of the nonlinear problem to a sequence of linear subproblems, enabling analytical or semi-analytical solutions in closed form.
- Rapid convergence demonstrated empirically in diverse contexts.
- Straightforward incorporation of multi-point (including nonlocal) boundary conditions [1310.2909].

**Limitations:**
- No adaptive mechanism for convergence control: for strongly nonlinear, stiff, or highly nonanalytic systems, the HPM series may lack sufficient convergence radius at $p=1$, necessitating use of HAM or introduction of auxiliary parameters [2604.13063].
- Sensitivity to the choice of linear operator $L$ and initial guess $u_0$; poor choices can impede convergence.
- Algebraic complexity increases for high-order terms or for strongly nonlinearities, especially in multi-variable settings.
- Rigorous convergence proofs available only in specific norms and under analytic/smallness assumptions [1310.2909].

**Potential Methodological Enhancements:**
- Integration with Padé approximants or Adomian decomposition to expand the region of convergence or accelerate series summation [1310.2909, 1507.02790].
- Use of hybrid transforms (Laplace, Natural, Sumudu) for memory/nonlocal/fractional differential equations [1611.06488, 1804.05735].
- Adoption of "optimal" variations (OHPM), introducing auxiliary functions or parameters to minimize global residuals and further improve convergence [1507.02790].
- Extension to multi-dimensional and strongly coupled systems via suitable operator and homotopy constructions.

## 7. Representative Case: Compact Stellar Objects via HPM

A concrete application is the use of HPM to solve the Tolman-Oppenheimer-Volkoff (TOV) equation for isotropic, spherically symmetric perfect fluids under a linear equation of state $p=\omega\rho$:
- The TOV equation is reduced to an ODE for the mass function $m(r)$ with nonlinear terms in $m,m',m''$ [1911.00350].
- The homotopy is constructed by separating the linear terms ($m'-\frac{1}{2}rm''$) from the nonlinear remainder.
- Power-series expansion with recursive determination of $m_0(r), m_1(r), m_2(r)$, etc., subject to regularity at $r=0$.
- The final approximate solution is a cubic (or higher) polynomial in $r$, with parameters determined by physical constraints (e.g., $p(R)=0$ for surface matching).
- The resulting models yield mass, compactness, redshift, and core-crust structure in agreement with known properties and numerically established maximum mass bounds [1911.00350].

This approach generalizes naturally to broader astrophysical, cosmological, and field-theory contexts, where high accuracy and analytic insight are desirable.

---

HPM thus provides a powerful, general, and conceptually unifying approach for the analysis and construction of analytical approximations to nonlinear differential equations across physics and applied mathematics. Its role as a special case of HAM contextualizes its strengths and boundaries, guiding rational application and motivating further methodological refinement.

Source: https://www.emergentmind.com/topics/homotopy-perturbation-method-hpm