Method of Riccati Pairs

• The method of Riccati pairs is a structured approach that pairs solutions of Riccati-type equations to restore closure and reveal underlying linearity.
• It acts as a regularization tool in quantum Darboux transformations by coupling Schrödinger solutions with Ermakov amplitudes to handle spectral and nodal challenges.
• Extensions into matrix analysis and control theory show that paired formulations yield effective addition laws and computational advantages in spectral splitting and stability analysis.

In the literature surveyed here, the method of Riccati pairs denotes a class of constructions in which the basic object is not an isolated Riccati solution but a paired structure that restores closure, exposes hidden linearity, or preserves spectral information. In one formulation, the pair consists of two solutions of the Riccati–Abel equation and their symmetric invariants; in another, it consists of a Schrödinger-generated Riccati quantity and an Ermakov amplitude; elsewhere it appears as a Riccati/dual-Riccati inequality pair, a pair of coupled algebraic Riccati equations, or a pair of matrices constrained by a Riccati inequality (Yamaleev, 2012, Blanco-Garcia et al., 2018, Kajántó, 2023). What is common to these uses is that pairing compensates for a structural deficiency of the single Riccati variable—typically the absence of an addition law, the loss of regularity, or the failure of direct linear closure.

1. Terminological scope and basic structural idea

The expression “method of Riccati pairs” is used in several related senses across the cited works. In the most explicit scalar setting, it refers to replacing a single solution $u(\phi)$ of the Riccati–Abel equation by a pair of solutions $(u,v)$, or equivalently by their symmetric combinations $t=-(u+v)$ and $s=uv$, because only this paired object admits a nontrivial summation law (Yamaleev, 2012). In the Darboux–Ermakov setting, it refers to pairing a Riccati superpotential with an Ermakov amplitude built from two linearly independent Schrödinger solutions, so that a complex Darboux transformation remains regular and spectrally controlled (Blanco-Garcia et al., 2018).

A broader reading is also present in later work. On model space forms, a Riccati pair is a pair $(L_\kappa,W)$ or $(L_\kappa,V)$ for which an admissible function satisfies a Riccati-type ordinary differential inequality, producing Hardy or Rellich inequalities (Kajántó, 2023). In diagonal Riccati stability, the “pair” is $(A,B)$, constrained by a Riccati inequality involving diagonal positive definite matrices $P,Q$ (Aleksandrov et al., 2015). In coupled algebraic Riccati theory, the relevant object is the pair $(X_1,X_2)$ of positive semidefinite solutions linked through Nash or Markov-jump couplings (Rajasingam et al., 2017, Li et al., 2020). In Riccati diagonalization of Hermitian matrices, the term is not used explicitly, but the construction encodes one invariant graph subspace and its complement through a matrix Riccati equation (Fujii et al., 2010).

This suggests that the phrase is best understood structurally rather than terminologically: the method operates by enlarging the state from one Riccati variable to a paired object on which the algebra, geometry, or spectral theory becomes tractable.

2. Riccati–Abel origin: pairing as a remedy for the missing addition law

The most direct formulation begins with the constant-coefficient Riccati equation

$u^2-a_1u+a_0=\frac{du}{d\phi},$

for which the usual single-variable addition formula exists: $(u,v)$0 This reflects the familiar relation between the ordinary Riccati equation, a second-order linear ODE, and a $(u,v)$1-dimensional complex algebra (Yamaleev, 2012).

For the cubic Riccati–Abel equation,

$(u,v)$2

the same program fails. Integration produces a logarithm of a product of three factors associated with the roots $(u,v)$3 of

$(u,v)$4

and the resulting algebra does not close on a single scalar solution. The paper states that there is no simple two-argument rational map $(u,v)$5 taking individual Riccati–Abel solutions to the solution at $(u,v)$6 in the same way as for the quadratic Riccati equation (Yamaleev, 2012).

The method of Riccati pairs resolves this by passing from a single solution to a pair. If $(u,v)$7 and $(u,v)$8 are solutions, the pair is encoded by the quadratic polynomial

$(u,v)$9

The main result is a composition law on these symmetric variables. If a second pair is encoded by $t=-(u+v)$0, then

$t=-(u+v)$1

with

$t=-(u+v)$2

and

$t=-(u+v)$3

This is Theorem 3.3 of the paper and is the precise summation formula for the Riccati–Abel equation: the operation closes on the pair $t=-(u+v)$4, not on a single $t=-(u+v)$5 (Yamaleev, 2012).

The significance of this construction is twofold. First, it identifies the obstruction in the cubic case as structural rather than accidental. Second, it supplies the additional degree of freedom needed for algebraic closure. The paper further states, conceptually, that for an $t=-(u+v)$6-th order generalized Riccati equation one needs an $t=-(u+v)$7-tuple of solutions—an “$t=-(u+v)$8-pulet”—to obtain an addition formula; in the cubic case this is exactly the Riccati pair (Yamaleev, 2012).

3. Linear and algebraic realization: companion matrices, multi-angle variables, and third-order complex algebra

The pair formulation for the Riccati–Abel equation is mirrored by a corresponding linear-algebraic structure. In the quadratic Riccati case, a $t=-(u+v)$9 matrix

$s=uv$0

satisfies

$s=uv$1

and the scalar Riccati solution is recovered as a ratio of components in the expansion of $s=uv$2. The cubic case requires a $s=uv$3 companion matrix

$s=uv$4

The exponential is no longer one-parameter; instead,

$s=uv$5

where $s=uv$6 satisfy coupled linear first-order systems (Yamaleev, 2012).

This two-parameter structure is essential. The paper shows that, under the constraint $s=uv$7 along a suitable curve $s=uv$8, the ratio

$s=uv$9

satisfies the Riccati–Abel equation. Unlike the ordinary Riccati case, a single scalar parameter is insufficient; the cubic case requires the pair $(L_\kappa,W)$0, and the nonlinear Riccati–Abel solution appears only after restriction to a codimension-one condition (Yamaleev, 2012).

The same section introduces tangent-type variables

$(L_\kappa,W)$1

with explicit addition formulas induced by the algebra of $(L_\kappa,W)$2. The Riccati-pair variables $(L_\kappa,W)$3 are the nonlinear projection of this third-order complex algebra. Thus, the method is not merely an ad hoc symmetrization; it is the nonlinear shadow of an underlying companion-matrix calculus and a generalized complex algebra of order three (Yamaleev, 2012).

4. Schrödinger–Ermakov–Darboux formulation

A second major use of the method appears in one-dimensional quantum mechanics, where the Riccati equation

$(L_\kappa,W)$4

arises from Darboux factorization of the stationary Schrödinger equation

$(L_\kappa,W)$5

Writing $(L_\kappa,W)$6, the imaginary part implies

$(L_\kappa,W)$7

and substitution yields the Ermakov equation

$(L_\kappa,W)$8

The Riccati pair arises by coupling a Schrödinger solution $(L_\kappa,W)$9 at energy $(L_\kappa,V)$0 with the Ermakov amplitude $(L_\kappa,V)$1 through the invariant

$(L_\kappa,V)$2

where $(L_\kappa,V)$3 (Blanco-Garcia et al., 2018).

For $(L_\kappa,V)$4, the amplitude can be written explicitly as

$(L_\kappa,V)$5

where $(L_\kappa,V)$6 and $(L_\kappa,V)$7 are linearly independent Schrödinger solutions with constant Wronskian $(L_\kappa,V)$8, and

$(L_\kappa,V)$9

This is the core structural statement: the Ermakov amplitude is a nonlinear superposition of two linearly independent Schrödinger solutions, and the complex Riccati superpotential

$(A,B)$0

is therefore a functional of the pair $(A,B)$1 (Blanco-Garcia et al., 2018).

The Darboux-transformed potential is

$(A,B)$2

Its discrete spectrum satisfies

$(A,B)$3

so the new non-Hermitian Hamiltonian inherits all discrete energies of the initial Hermitian system and gains one additional real eigenvalue $(A,B)$4. The paper shows that $(A,B)$5 can remain regular even when $(A,B)$6 coincides with a discrete energy or lies between two discrete energies, because $(A,B)$7 has no zeros in $(A,B)$8 although $(A,B)$9 and $P,Q$0 separately may have zeros (Blanco-Garcia et al., 2018).

In this setting, the method of Riccati pairs is therefore a regularization mechanism. The pair $P,Q$1 compensates for the nodal singularities that would obstruct a conventional one-step real Darboux transformation, and it simultaneously controls the gain–loss structure

$P,Q$2

under the stated boundary behavior (Blanco-Garcia et al., 2018).

5. Variational and inequality-theoretic extensions

The pairing idea also enters in odd-order Riccati chains. For

$P,Q$3

the Cole–Hopf substitution

$P,Q$4

maps the $P,Q$5-th Riccati member to a linear ODE of order $P,Q$6: $P,Q$7 For odd $P,Q$8, the associated linear equation is even order, can be brought to self-adjoint form, and admits a standard Lagrangian built from the self-adjoint operator. The resulting Riccati Lagrangians are linear in the highest derivative, and the paper concludes that the Riccati family cannot be Hamiltonized by the traditional Legendre-transform method used in classical mechanics (Sarkar et al., 2024).

On model space forms, the method appears in a distinct but related way. A pair $P,Q$9 is a $(X_1,X_2)$0-Riccati pair if there exists $(X_1,X_2)$1 such that

$(X_1,X_2)$2

while $(X_1,X_2)$3 is a $(X_1,X_2)$4-dual Riccati pair if there exists $(X_1,X_2)$5 such that

$(X_1,X_2)$6

These inequalities yield Hardy- and Rellich-type estimates after convexity arguments and integration by parts. In Euclidean space the dual-pair construction recovers the classical Rellich inequality

$(X_1,X_2)$7

by composing a Rellich inequality for $(X_1,X_2)$8 with a Hardy inequality generated by a Riccati pair (Kajántó, 2023).

These developments preserve the same logic as the Riccati–Abel and Darboux–Ermakov cases. The single equation is replaced by a structurally compatible partner—linear, dual, or self-adjoint—so that one gains access to addition laws, variational formulas, or sharp integral inequalities that are not directly visible at the single-Riccati level.

6. Matrix, control-theoretic, and numerical formulations

In matrix analysis and control, the language of Riccati pairs shifts from scalar solutions to structured pairs of matrices. For real matrices $(X_1,X_2)$9, the pair $u^2-a_1u+a_0=\frac{du}{d\phi},$0 is called diagonally Riccati stable if there exist diagonal positive definite matrices $u^2-a_1u+a_0=\frac{du}{d\phi},$1 such that

$u^2-a_1u+a_0=\frac{du}{d\phi},$2

A necessary and sufficient condition is given in terms of the sign of diagonal entries of $u^2-a_1u+a_0=\frac{du}{d\phi},$3 for every nonzero positive semidefinite block matrix

$u^2-a_1u+a_0=\frac{du}{d\phi},$4

with $u^2-a_1u+a_0=\frac{du}{d\phi},$5. For Metzler $u^2-a_1u+a_0=\frac{du}{d\phi},$6 and nonnegative $u^2-a_1u+a_0=\frac{du}{d\phi},$7, diagonal Riccati stability is equivalent to $u^2-a_1u+a_0=\frac{du}{d\phi},$8 being Hurwitz, and this criterion is then used to construct Lyapunov–Krasovskii functionals for generalized Lotka–Volterra systems with delay (Aleksandrov et al., 2015).

In coupled algebraic Riccati theory, the relevant pair is $u^2-a_1u+a_0=\frac{du}{d\phi},$9. For the continuous coupled algebraic Riccati equation

$(u,v)$00

the accelerated Riccati iteration produces monotone and bounded sequences under explicit order assumptions. In the increasing case it converges to the minimal positive semidefinite solution; in the decreasing case it converges to the maximal one. The paper further proves that the accelerated iteration is faster, in the Loewner-order sense, than the regular iteration (Rajasingam et al., 2017). A related two-player Nash framework treats coupled AREs through the pair $(u,v)$01 together with feedback matrices $(u,v)$02, iteratively decoupling the problem into ordinary AREs for each player (Li et al., 2020).

Riccati diagonalization of Hermitian matrices provides a geometric counterpart. For a block Hermitian matrix

$(u,v)$03

one introduces a graph parameter $(u,v)$04 and a unitary $(u,v)$05 so that block-diagonalization is achieved precisely when $(u,v)$06 satisfies the matrix Riccati equation

$(u,v)$07

The graph subspace $(u,v)$08 and its orthogonal complement then form the paired invariant structures underlying the decomposition (Fujii et al., 2010).

Taken together, these matrix formulations show that the method of Riccati pairs is not confined to scalar differential equations. It extends to invariant subspaces, diagonal stability, spectral splitting, and monotone computation of extremal solutions. A plausible implication is that the unifying content of the method lies in pairing complementary objects—solutions, subspaces, amplitudes, inequalities, or feedback matrices—until the Riccati structure becomes closed under the operation one wishes to study.