1 Definition and basic idea

A similarity solution is a special type of solution to a differential equation in which the dependent variables depend on the independent variables only through a reduced combination called a similarity variable. This reduction collapses the number of free variables and often turns a partial differential equation into an ordinary differential equation. The resulting form is typically easier to analyze, and in some cases it can be solved exactly.

Similarity solutions appear when a problem contains a scaling structure or an invariance under rescaling of time, space, or other variables. They are especially useful in systems where the overall shape of the solution is preserved while its size, amplitude, or characteristic length changes.

1.1 Concept of self-similarity

Self-similarity refers to the property that a pattern looks the same at different scales. In mathematics, this means that a solution can be rescaled without changing its essential form. A self-similar solution may describe a spreading diffusion profile, a thinning boundary layer, or a wave whose shape remains fixed after appropriate normalization.

Self-similar behavior can occur exactly or approximately. Exact self-similarity arises when the governing equations and boundary conditions are invariant under a scaling transformation. Approximate self-similarity often appears in long-time or short-time limits, where the solution approaches a universal form.

1.2 Similarity variables

A similarity variable is a combined coordinate built from the original independent variables so that the solution depends on fewer arguments. For example, a variable of the form x/t^a or x/\sqrt{t} combines space and time into a single dimensionless quantity. The choice of similarity variable is dictated by the symmetry or scaling properties of the problem.

Once the similarity variable is introduced, the dependent variable is typically written as a product of a scaling factor and an unknown function of the reduced variable. This ansatz captures how the amplitude and shape evolve together.

1.3 Reduction of differential equations

The main advantage of similarity methods is the reduction of complexity. A PDE with several variables may become an ODE in one similarity variable, or a system of ODEs if multiple dependent fields are involved. This reduction can expose hidden structure and make numerical or analytical treatment more straightforward.

1.3.1 From partial differential equations to ordinary differential equations

A PDE is reduced by substituting the similarity form into the original equation and using the chain rule to rewrite derivatives in terms of the reduced variable. The PDE then collapses to an ODE for the similarity profile. In many classical problems, this ODE has boundary conditions at finite or infinite values of the similarity variable.

1.3.2 Role of scaling transformations

Scaling transformations describe how variables change under multiplication by powers of a scale factor. If the equation remains unchanged under such transformations, the problem may admit a similarity solution. These transformations often determine the powers appearing in the similarity variable and in the amplitude factor.

1.4 Types of similarity solutions

Similarity solutions are commonly grouped by the kind of scaling they use. First-kind similarity solutions are determined directly by dimensional analysis and symmetry. Second-kind similarity solutions involve scaling exponents that are not fixed by dimensions alone and must be found from an eigenvalue-like condition. In some contexts, intermediate or dynamic similarity appears when a solution approaches a self-similar form over time without being exactly invariant.

2 Mathematical formulation

The mathematical basis of similarity solutions lies in the presence of invariance, reduced variables, and dimensionally consistent scaling. The method is usually organized by identifying a transformation group, constructing invariants, and rewriting the dependent variables in terms of those invariants.

2.1 Dimensional analysis

Dimensional analysis helps identify combinations of variables that are dimensionless or have the same units as the dependent variable. It is often the first step in finding a similarity form. By balancing dimensions, one can determine the exponents in a self-similar ansatz and isolate the most natural reduced coordinate.

This approach is especially effective when the governing equation contains only a few physical parameters. It can reveal characteristic length scales, time scales, and velocity scales that determine the structure of the similarity solution.

2.2 Scaling symmetries

Scaling symmetries are transformations that multiply variables by powers of a common factor while leaving the governing equation unchanged. When such a symmetry exists, the equation may be reducible to an invariant form. The exponents in the transformation determine how the dependent and independent variables should be rescaled.

These symmetries are central to the construction of similarity solutions because they encode the relation between different scales in the problem. They also help classify solutions according to the type of invariance they satisfy.

2.3 Invariant transformations

An invariant transformation is a change of variables that leaves the essential form of the equation unchanged. Under such a transformation, certain combinations of the original variables remain fixed. These invariant combinations become the similarity variables used in the reduced formulation.

Invariant transformations can arise from continuous symmetry groups. In such cases, the solution is built from group invariants and can often be interpreted as a fixed point under the action of the symmetry.

2.4 Construction of similarity ansatz

The similarity ansatz is the assumed functional form used to reduce the equation. It typically expresses the solution as a product or composition of scaling factors and a reduced profile function. The ansatz is chosen to match the problem’s symmetry, boundary conditions, and physical dimensions.

2.4.1 Choice of dependent variables

The dependent variables are written in a way that separates scaling from shape. For example, temperature, velocity, or concentration may be expressed as a prefactor times an unknown profile of the similarity variable. The prefactor accounts for amplitude change, while the profile describes the invariant shape.

2.4.2 Choice of independent variables

The independent variables are combined into one or more similarity variables that reflect the symmetry of the problem. The choice must preserve the relevant boundary conditions and reduce the number of independent coordinates without losing essential physical information.

3 Methods of obtaining similarity solutions

Several complementary methods are used to derive similarity solutions. Some rely on direct guesswork guided by physical insight, while others use systematic symmetry analysis or asymptotic reasoning. Numerical methods are often employed when the reduced equation remains nonlinear or difficult to solve.

3.1 Direct substitution

Direct substitution is the simplest method. One proposes a similarity form, inserts it into the governing equation, and checks whether the equation reduces consistently. If successful, the resulting reduced equation provides the similarity profile.

This method is effective when the expected scaling can be inferred from dimensional arguments or physical intuition. It is widely used in classical problems such as diffusion, boundary layers, and source-driven spreading.

3.2 Lie group analysis

Lie group analysis provides a systematic framework for finding symmetries of differential equations. It searches for continuous transformation groups under which the equation is invariant. Once these symmetries are known, they can be used to construct similarity variables and reduce the equation.

3.2.1 Infinitesimal generators

Infinitesimal generators describe the local action of a symmetry group. They encode how each variable changes under a small transformation. By solving the associated determining equations, one can identify the symmetries admitted by the differential equation.

3.2.2 Invariants and reduction

The invariants of the symmetry group are combinations of variables unchanged by the transformation. These invariants serve as the reduced independent variables in the similarity solution. The dependent variables are then expressed in terms of these invariants, yielding a lower-dimensional system.

3.3 Dimensional and asymptotic methods

Dimensional methods use units and scale arguments to predict the form of the similarity variable. Asymptotic methods examine limiting regimes, such as large time or small distance, where the solution may approach a universal profile. Together, these methods often suggest the correct similarity form before any detailed calculation is done.

3.4 Numerical approaches

When the reduced ODE cannot be solved in closed form, numerical integration is commonly used. Similarity reduction is still valuable because it lowers the computational dimension and simplifies boundary conditions. Numerical similarity profiles can also be compared across parameter values to study universality and transition between regimes.

4 Applications in science and engineering

Similarity solutions are widely used in fields where transport, diffusion, and flow exhibit scaling behavior. They provide benchmark solutions, illuminate dominant balances, and guide approximate modeling.

4.1 Fluid mechanics

In fluid mechanics, similarity methods are especially important for laminar flow, viscous layers, and free-shear flows. They help describe velocity profiles and characteristic thickness growth in situations where the governing equations are nonlinear.

4.1.1 Boundary-layer flows

Boundary-layer theory often leads to similarity reductions because the flow near a surface has a natural scaling in the streamwise and normal directions. The resulting profiles describe how velocity changes across a thin region adjacent to the boundary. Classical examples include flow over a flat plate and related shear-layer problems.

4.1.2 Jets and wakes

Jets and wakes frequently develop self-similar shapes as they spread downstream. The velocity field can often be expressed in terms of a scaled transverse coordinate. In such cases, the width grows while the peak intensity decreases in a predictable way.

4.2 Heat transfer and diffusion

Heat and mass transport problems are among the most familiar settings for similarity solutions. Because diffusion generates characteristic spreading laws, many solutions naturally depend on x/\sqrt{t} or related variables.

4.2.1 Heat conduction

The heat equation admits classic similarity solutions that describe the spreading of thermal disturbances. These solutions often involve error-function-like profiles and capture the gradual smoothing of temperature gradients over time.

4.2.2 Mass diffusion

Diffusion problems for concentration fields can be treated in the same way as heat conduction. Similarity forms are used for point release, semi-infinite media, and diffusion fronts. The reduced profile typically describes a smooth transition between high and low concentration regions.

4.3 Wave phenomena

Similarity ideas also appear in wave propagation, especially when a wavefront expands with a characteristic scaling law. In such cases, the field may depend on a radius-time combination that captures the geometry of the spread. Similarity methods are useful in studying blast-like expansions, nonlinear wave fronts, and asymptotic dispersive behavior.

4.4 Astrophysics and plasma physics

In astrophysics and plasma physics, self-similar models are used to describe expanding gases, shock-like structures, and large-scale transport processes. They are valuable when the system exhibits broad scale separation and the exact solution is not available. Similarity forms can simplify the description of evolving density, pressure, or magnetic-field distributions.

4.5 Chemical reaction and combustion problems

Reaction and combustion systems may admit similarity solutions when reaction rates, diffusion, and heat release balance in a scale-invariant way. These solutions are useful for flame structure, ignition fronts, and spreading reaction zones. They can clarify how chemical and transport effects interact across evolving length scales.

5 Examples of similarity solutions

Several classical examples illustrate how similarity solutions work in practice. These cases are widely used in textbooks and serve as benchmarks for more complicated models.

5.1 Blasius boundary-layer solution

The Blasius solution describes steady laminar flow over a flat plate. By introducing a similarity variable based on the transverse coordinate and the streamwise distance, the boundary-layer PDEs reduce to a nonlinear ODE. The resulting velocity profile is a foundational example in viscous flow theory.

5.2 Heat equation similarity solution

For the one-dimensional heat equation, a common similarity variable is x/\sqrt{t}. This reduction yields a profile that captures the smoothing of an initially localized thermal disturbance. The solution often involves the error function and illustrates the universal spreading of diffusion.

5.3 Diffusion from a point source

A point-source diffusion problem generates a radially symmetric self-similar profile. The distribution spreads outward while its peak decreases so that total mass remains conserved. The resulting form depends on radius and time through a reduced scaling combination.

5.4 Burgers' equation reductions

Burgers' equation combines nonlinear advection and diffusion, making it a useful model for shock formation and smoothing. Under appropriate transformations, it can admit similarity solutions describing evolving layers or decaying pulses. These reductions are important in the study of nonlinear transport and model turbulence.

6 Properties and interpretation

Similarity solutions are more than algebraic conveniences. They encode physical structure, reveal conservation laws in disguised form, and often identify universal behavior shared across different systems.

6.1 Scaling behavior

The most visible property of a similarity solution is its scaling law. As the independent variables change, the solution changes in a coordinated way that preserves its overall form. This behavior often determines how width, amplitude, or front location evolves with time or distance.

6.2 Stability and validity

A similarity solution is useful only within its domain of validity. Some are exact solutions of the governing equations, while others approximate the behavior of a system over a limited range. Stability also matters, since a similarity form may describe the observed profile only if perturbations decay or remain bounded.

6.3 Exact versus approximate similarity

Exact similarity arises directly from invariance of the equations and boundary conditions. Approximate similarity appears when the solution approaches a scaled form asymptotically, often after transients have faded. Both types are important, but approximate forms usually require careful justification.

6.4 Physical meaning of reduced variables

The reduced variables condense the essential physics into a compact form. A similarity variable often measures distance relative to a characteristic spreading length, while a reduced profile describes how the field varies across that scaled coordinate. This interpretation makes similarity solutions especially useful for comparing different systems on a common basis.

Similarity methods are connected to a broader set of ideas in differential equations, asymptotic analysis, and dynamical systems. These related concepts help explain why self-similar forms emerge and how they evolve.

7.1 Intermediate asymptotics

Intermediate asymptotics refers to a regime where the solution has forgotten many details of the initial or boundary data but has not yet reached a final steady state. In this stage, a universal similarity form may dominate. Such behavior is common in diffusion and spreading phenomena.

7.2 Traveling wave solutions

Traveling wave solutions are related but distinct. Instead of scaling in place, they maintain a fixed shape while moving at constant speed. Some equations admit both traveling and self-similar forms, depending on the balance of transport, diffusion, and reaction effects.

7.3 Lie symmetries

Lie symmetries provide the structural framework underlying many similarity reductions. They identify continuous invariances of differential equations and supply the invariants used in reduction. The theory is a powerful tool for classification and systematic solution generation.

7.4 Self-similar attractors

A self-similar attractor is a solution profile approached by a wide class of initial conditions after suitable rescaling. In this sense, the similarity form acts as an organizing state for the dynamics. Such attractors are important in nonlinear evolution equations because they explain why diverse starting configurations can converge to the same scaled pattern.