Analytical solutions of the Riccati equation with coefficients satisfying integral or differential conditions with arbitrary functions
aa r X i v : . [ m a t h . C A ] N ov Analytical solutions of the Riccati equation withcoefficients satisfying integral or differentialconditions with arbitrary functions
Tiberiu Harko , ∗ , Francisco S. N. Lobo , M. K. Mak Department of Mathematics, University College London, Gower Street, London WC1E 6BT, United Kingdom Centro de Astronomia e Astrof´ısica da Universidade de Lisboa, Campo Grande, Ed. C8 1749-016 Lisboa, Portugal Department of Computing and Information Management, Hong Kong Institute of Vocational Education, Chai Wan, Hong Kong, P. R. China ∗ Corresponding Author: [email protected]
Abstract
Ten new exact solutions of the Riccati equation dy/dx = a ( x ) + b ( x ) y + c ( x ) y are presented. Thesolutions are obtained by assuming certain relations among the coefficients a ( x ), b ( x ) and c ( x ) of the Riccatiequation, in the form of some integral or differential expressions, also involving some arbitrary functions. By ap-propriately choosing the form of the coefficients of the Riccati equation, with the help of the conditions imposed onthe coefficients, we obtain ten new integrability cases for the Riccati equation. For each case the general solutionof the Riccati equation is also presented. The possibility of the application of the obtained mathematical resultsfor the study of anisotropic general relativistic stellar models is also briefly considered. Keywords: Riccati equation; integrability condition; exact solutions: anisotropic rel-ativistic stellar models; stars
The Riccati equation is one of the most studied first order non-linear differential equations [1, 2, 3], and is givenby dydx = a ( x ) + b ( x ) y + c ( x ) y , (1)where a , b , c are arbitrary real functions of x , with a, b, c ∈ C ∞ ( I ), defined on a real interval I ⊆ ℜ .It is well-known that once a particular solution y p of the Riccati equation is known, the general solution ofEq. (1) is given by y [ b ( x ) , c ( x ) , y p ( x )] = y p ( x ) + e R x [ b ( φ )+2 c ( φ ) y p ( φ )] dφ C − R x c ( ψ ) e R ψ [ b ( φ )+2 c ( φ ) y p ( φ )] dφ dψ , (2)where C is an arbitrary constant of integration and the particular solution y p ( x ) satisfies the Riccati equation dy p dx = a ( x ) + b ( x ) y p ( x ) + c ( x ) y p ( x ) . (3)If we know three particular solutions y pi ( x ), i = 1 , ,
3, then the Riccati equation can be solved without quadratures[1, 2, 3].Even when a particular solution is not known, the Riccati equation can be integrated exactly if the coefficientsof the equation satisfy some conditions. For instance, if the coefficients of the Riccati equation satisfy the followingspecific condition a ( x ) + b ( x ) + c ( x ) ≡ , (4)then the solution of the Riccati equation is given by y = K + R [ c ( x ) + a ( x )] E ( x ) dx − E ( x ) K + R [ c ( x ) + a ( x )] E ( x ) dx + E ( x ) , (5)where K is an arbitrary constant of integration, and E ( x ) = exp( R [ c ( x ) − a ( x )] dx [1].If the coefficients of the Riccati equation satisfy the more general condition λ c ( x ) + λµb ( x ) + µ a ( x ) ≡ , (6)1here λ and µ are arbitrary constants satisfying the condition | λ | + | µ | >
0, then by means of the transformation y ( x ) = λ/µ + u ( x ), the Riccati equation is transformed into a Bernoulli type equation [1], u ′ ( x ) = c ( x ) u ( x ) + (cid:20) λµ c ( x ) + b ( x ) (cid:21) u ( x ) . (7)Another interesting case is if the coefficients a ( x ), b ( x ), c ( x ) satisfy the relation a ( x ) + b ( x ) + c ( x ) = ddx ln (cid:12)(cid:12)(cid:12)(cid:12) α ( x ) β ( x ) (cid:12)(cid:12)(cid:12)(cid:12) − α ( x ) − β ( x ) α ( x ) β ( x ) [ α ( x ) c ( x ) − β ( x ) a ( x )] , (8)with α ( x ) and β ( x ) properly chosen differentiable functions, such that αβ >
0, then the Riccati equation isintegrable by quadratures [4]. If c ( x ) ≡
1, and the functions a ( x ) and b ( x ) are polynomials satisfying the condition∆ = b ( x ) − db ( x ) dx − a ( x ) ≡ constant , (9)then y ± ( x ) = − h b ( x ) ± √ ∆ i , (10)are both solutions of the Riccati Eq. (1) [1, 3, 5]. Very recently, the integrability condition given by Eq. (9) of theRiccati Eq. (1), and of the reduced Riccati equation of the form dy/dx = a ( x ) + c ( x ) y , have been generalized in[6, 7].Note that the Riccati equation plays a significant role in many fields of applied and fundamental science [8].Some applications of the integrability conditions for the case of the damped harmonic oscillator with time dependentfrequency and for solitonic wave have been briefly discussed in [6]. The applications of the integrability conditionof the reduced Riccati equation for the integration of the Schr¨odinger and Navier-Stokes equations have been alsostudied in [7]. The integrability conditions derived in [6, 7] have been applied in [9] to obtain a general solution ofthe Einstein’s gravitational field equations for the static spherically symmetric gravitational interior space-time ofan isotropic fluid sphere. The astrophysical analysis indicates that this solution can be used as a realistic model forstatic general relativistic high density objects, such as neutron stars. Riccati equations also play an important rolein cosmology. The mechanism of the initial inflationary scenario of the Universe and of its late-time accelerationcan be described by assuming the existence of some gravitationally coupled scalar fields φ , with the inflaton fieldgenerating inflation and the quintessence field being responsible for the late accelerated expansion of the Universe.In the case of a scalar field dominated Universe the dynamics of the expansion can be described by the solutionsof a Riccati type equation [10]. In fact, the integrability conditions of the Riccati equation, given by Eq. (1), andits applications to stellar and cosmological models have been extensively discussed in the literature, and we referthe reader to [11, 12, 13, 14].Due to the nonlinear structure of the Riccati equation, the general solution of the Riccati equation (1), cannotbe easily found. Therefore, one has to use numerical techniques, or approximate method for obtaining its solutions.Recently, a numerical method, using hybrid of block-pulse functions, and Chebyshev polynomials for solving theRiccati equation has been presented in [15]. Various numerical methods such as Adomian decomposition method,He’s variational iteration method, homopoty perturbation method, Taylor matrix method and Legendre waveletmethod for solving the Riccati equation have been proposed in [16, 17, 18, 19, 20, 21, 22].It is the purpose of the present paper to present some further integrability conditions of the Riccati equation,given by Eq. (1). The relations among the coefficients of the Riccati equation involve some integral or differentialrepresentations, as well as the presence of some arbitrary functions. By appropriately choosing the form of thecoefficients of the Riccati equation we obtain ten new integrability conditions. For each case the general solutionof the Riccati equation is also obtained. The possibility of the application of the obtained mathematical results forthe study of the anisotropic general relativistic stellar models is also briefly considered.The present paper is organized as follows. The ten new exact solutions of the Riccati Eq. (1) are presented inSection 2. Some astrophysical applications of the solutions of the Riccati equation are presented in Section 3. Weconclude our results in Section 4. From the algebraic point of view the Riccati Eq. (1) is a quadratic equation in y ( x ). We consider that itsparticular solutions y p ± ( x ) take the form y p ± ( x ) = − b ( x ) ± q b ( x ) − a ( x ) c ( x ) + 4 c ( x ) dy p dx c ( x ) . (11)2n order to obtain the general solution of the Riccati Eq. (1) with the help of Eq. (11), we introduce the newgenerating function f ( x ), satisfying the differential condition given by b ( x ) + 4 c ( x ) dy p dx = f ( x ) , (12)representing a first order differential equation in y p ( x ), and which can be immediately integrated to give theparticular solution of the Riccati Eq. (1). Therefore the general solution of the Riccati Eq. (1) can be obtainedthrough quadratures, since its particular solution is known.Based on the above algorithm, we shall show the detailed calculations in obtaining Theorem 1 in Section 2.1,case 1. In order to make the paper readable and not to have repetitive calculations, we shall not present thedetailed calculations for the rest of the theorems here. However, one may follow the same procedures to obtain theTheorems 2-9 presented in this paper. In this Section, we shall present ten new integrability cases of the RiccatiEq. (1). For each case the general solution of the Riccati equation is explicitly obtained. a ( x ) = f ( x ) − (cid:26) b ( x )+ c ( x ) (cid:20)R x f φ ) − b φ )2 c ( φ ) dφ − C (cid:21)(cid:27) c ( x ) We assume that the arbitrary functions b ( x ), c ( x ) and f ( x ) satisfy the differential condition b ( x ) + 4 c ( x ) dy p dx = f ( x ) , (13)where we have introduced a generating function f ( x ) ∈ C ∞ ( I ) defined on a real interval I ⊆ ℜ . The particularsolutions y p ± ( x ) of the Riccati Eq. (1) take the form y p ± ( x ) = − b ( x ) ± p f ( x ) − a ( x ) c ( x )2 c ( x ) = 12 (cid:20)Z x f ( φ ) − b ( φ )2 c ( φ ) dφ − C (cid:21) . (14)where we have used and integrated Eq. (13), and C is an arbitrary constant of integration. By differentiatingEq. (14) with respect to x yields the result ddx " − b ( x ) ± p f ( x ) − a ( x ) c ( x ) c ( x ) = f ( x ) − b ( x )2 c ( x ) . (15)Eq. (15) can be integrated to give the coefficient a ( x ) of the Riccati Eq. (1) as a ( x ) = f ( x ) − n b ( x ) + c ( x ) hR x f ( φ ) − b ( φ )2 c ( φ ) dφ − C io c ( x ) , (16)where C is an arbitrary constant of integration. By substituting Eq. (16) into the Riccati Eq. (1), the latter canbe expressed as dydx = f ( x ) − n b ( x ) + c ( x ) hR x f ( φ ) − b ( φ )2 c ( φ ) dφ − C io c ( x ) + b ( x ) y + c ( x ) y . (17)Therefore we obtain the following: Theorem 1.
If the coefficient a ( x ) of the Riccati Eq. (1) satisfies the integral condition (16), then the generalsolution of the Riccati Eq. (17) is given by y ( x ) = e R x (cid:26) b ( φ )+ c ( φ ) (cid:20)R φ f ψ ) − b ψ )2 c ( ψ ) dψ − C (cid:21)(cid:27) dφ C − R x c ( ψ ) e R ψ n b ( φ )+ c ( φ ) hR φ f ψ ) − b ψ )2 c ( ψ ) dψ − C io dφ dψ + 12 (cid:20)Z x f ( φ ) − b ( φ )2 c ( φ ) dφ − C (cid:21) , (18)where C is an arbitrary constant of integration. a ( x ) = ddx (cid:20) − b ( x ) ± √ f ( x )+ b ( x )2 c ( x ) (cid:21) − f ( x )4 c ( x ) Next, we assume that the arbitrary functions a ( x ) satisfies the differential condition a ( x ) = ddx " − b ( x ) ± p f ( x ) + b ( x )2 c ( x ) − f ( x )4 c ( x ) , (19)3here we have introduced a new arbitrary function f ( x ) ∈ C ∞ ( I ) defined on a real interval I ⊆ ℜ . By substitutingEq. (19) into the Riccati Eq. (1), the latter can be expressed as dy ± dx = ( ddx " − b ( x ) ± p f ( x ) + b ( x )2 c ( x ) − f ( x )4 c ( x ) ) + b ( x ) y ± + c ( x ) y ± . (20)Therefore we obtain the following: Theorem 2 . If the coefficient a ( x ) of the Riccati Eq. (1) satisfies the differential condition (19), then thegeneral solutions of the Riccati Eq. (20) are given by y ± ( x ) = e ± R x √ f ( φ )+ b ( φ ) dφ C ± − R x c ( ψ ) e ± R ψ √ f ( φ )+ b ( φ ) dφ dψ + " − b ( x ) ± p f ( x ) + b ( x )2 c ( x ) , (21)where C ± are arbitrary constants of integration. b ( x ) = f ( x ) − c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o We now assume that the arbitrary function b ( x ) satisfies the integral condition b ( x ) = f ( x ) − c ( x ) nR x h a ( φ ) + f ( φ )4 c ( φ ) i dφ − C o c ( x ) nR x h a ( φ ) + f ( φ )4 c ( φ ) i dφ − C o , (22)where C is an arbitrary constant. By substituting Eq. (22) into the Riccati Eq. (1), the latter can be expressed as dydx = a ( x ) + f ( x ) − c ( x ) nR x h a ( φ ) + f ( φ )4 c ( φ ) i dφ − C o c ( x ) nR x h a ( φ ) + f ( φ )4 c ( φ ) i dφ − C o y + c ( x ) y . (23)Therefore we obtain the following: Theorem 3 . If the coefficient b ( x ) of the Riccati Eq. (1) satisfies the integral condition (22), then the generalsolution of the Riccati Eq. (23) is given by y ( x ) = e − R x vuuut f ( ω )+ f ω ) − c ω ) (cid:26)R ω (cid:20) a ( φ )+ f φ )4 c ( φ ) (cid:21) dφ − C (cid:27) c ( ω ) (cid:26)R ω (cid:20) a ( φ )+ f φ )4 c ( φ ) (cid:21) dφ − C (cid:27) dω C − R x c ( ψ ) e − R ψ vuuut f ( ω )+ f ω ) − c ω ) (cid:26)R ω (cid:20) a ( φ )+ f φ )4 c ( φ ) (cid:21) dφ − C (cid:27) c ( ω ) (cid:26)R ω (cid:20) a ( φ )+ f φ )4 c ( φ ) (cid:21) dφ − C (cid:27) dω dψ − c ( x ) vuut f ( x ) + ( f ( x ) − c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o ) + f ( x ) − c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o c ( x ) nR x h a ( φ )+ f φ )4 c ( φ ) i dφ − C o , (24)where C is an arbitrary constant of integration. c ( x ) = h − b ( x ) ± √ f ( x )+ b ( x ) i e − R x f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ C ± + R x a ( ψ ) e − R ψ f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ dψ We assume that the arbitrary function c ( x ) satisfies the integral condition c ( x ) = h − b ( x ) ± p f ( x ) + b ( x ) i e − R x f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ (cid:20) C ± + R x a ( ψ ) e − R ψ f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ dψ (cid:21) , (25)4here C ± are arbitrary constants. By substituting Eq. (25) into the Riccati Eq. (1), the latter can be expressedas dy ± dx = a ( x ) + b ( x ) y ± + h − b ( x ) ± p f ( x ) + b ( x ) i e − R x f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ (cid:20) C ± + R x a ( ψ ) e − R ψ f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ dψ (cid:21) y ± . (26)Therefore we obtain the following: Theorem 4 . If the coefficient c ( x ) of the Riccati Eq. (1) satisfies the integral condition (25), then the generalsolutions of the Riccati Eq. (26) are given by y ± ( x ) = e ± R x √ f ( φ )+ b ( φ ) dφ C ± − R x h − b ( ω ) ± √ f ( ω )+ b ( ω ) i e − R ω f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ C ± + R ω a ( ψ ) e − R ψ f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ dψ e ± R ω √ f ( φ )+ b ( φ ) dφ dω + (cid:20) C ± + Z x a ( ψ ) e − R ψ f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ dψ (cid:21) e R x f φ ) − b ( φ ) ± √ f φ )+ b φ ) dφ , (27)where C ± are arbitrary constants of integration. a ( x ) = (cid:26) f ( x ) c ( x ) − b ( x ) hR x f ( φ )2 c ( φ ) dφ − C i − c ( x ) hR x f ( φ )2 c ( φ ) dφ − C i (cid:27) Assume now that the arbitrary function a ( x ) satisfies the integral condition a ( x ) = 14 ( f ( x ) c ( x ) − b ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21) − c ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21) ) , (28)where f ( x ) ∈ C ∞ ( I ) is an arbitrary function defined on a real interval I ⊆ ℜ and C is an arbitrary constant.By substituting Eq. (28) into the Riccati Eq. (1), the latter can be expressed as dydx = 14 ( f ( x ) c ( x ) − b ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21) − c ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21) ) + b ( x ) y + c ( x ) y . (29)Therefore we obtain the following: Theorem 5 . If the coefficient a ( x ) of the Riccati Eq. (1) satisfies the integral condition (28), then the generalsolution of the Riccati Eq. (29) is given by y ( x ) = e R x r b ( ψ )+ c ( ψ ) hR ψ f φ )2 c ( φ ) dφ − C in b ( ψ )+ c ( ψ ) hR ψ f φ )2 c ( φ ) dφ − C io dψ C − R x c ( ω ) e R ω r b ( ψ )+ c ( ψ ) hR ψ f φ )2 c ( φ ) dφ − C in b ( ψ )+ c ( ψ ) hR ψ f φ )2 c ( φ ) dφ − C io dψ dω + 12 c ( x ) ( − b ( x ) + s b ( x ) + c ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21) (cid:26) b ( x ) + c ( x ) (cid:20)Z x f ( φ )2 c ( φ ) dφ − C (cid:21)(cid:27)) , (30)where C is an arbitrary constant of integration. b ( x ) = f ( x ) − a ( x ) c ( x ) − c ( x ) hR x f φ )2 c ( φ ) dφ − C i c ( x ) hR x f φ )2 c ( φ ) dφ − C i In this case, we assume that the coefficient b ( x ) of the Riccati Eq. (1) satisfies the integral condition b ( x ) = f ( x ) − a ( x ) c ( x ) − c ( x ) hR x f ( φ )2 c ( φ ) dφ − C i c ( x ) hR x f ( φ )2 c ( φ ) dφ − C i . (31)By substituting Eq. (31) into the Riccati Eq. (1), the latter can be expressed as dydx = a ( x ) + f ( x ) − a ( x ) c ( x ) − c ( x ) hR x f ( φ )2 c ( φ ) dφ − C i c ( x ) hR x f ( φ )2 c ( φ ) dφ − C i y + c ( x ) y . (32)5herefore we obtain the following: Theorem 6 . If the coefficient b ( x ) of the Riccati Eq. (1) satisfies the integral condition (31), then the generalsolution of the Riccati Eq. (32) is given by y ( x ) = e − R x vuuut f ψ ) − a ( ψ ) c ( ψ ) − c ψ ) (cid:20)R ψ f φ )2 c ( φ ) dφ − C (cid:21) c ( ψ ) (cid:20)R ψ f φ )2 c ( φ ) dφ − C (cid:21) − a ( ψ ) c ( ψ )+ f ( ψ ) dψ C − R x c ( ω ) e − R ω vuuut f ψ ) − a ( ψ ) c ( ψ ) − c ψ ) (cid:20)R ψ f φ )2 c ( φ ) dφ − C (cid:21) c ( ψ ) (cid:20)R ψ f φ )2 c ( φ ) dφ − C (cid:21) − a ( ψ ) c ( ψ )+ f ( ψ ) dψ dω + 12 c ( x ) − vuut( f ( x ) − a ( x ) c ( x ) − c ( x ) hR x f φ )2 c ( φ ) dφ − C i c ( x ) hR x f φ )2 c ( φ ) dφ − C i ) − a ( x ) c ( x ) + f ( x )+ a ( x ) c ( x ) − f ( x )+ c ( x ) hR x f φ )2 c ( φ ) dφ − C i c ( x ) hR x f φ )2 c ( φ ) dφ − C i , (33)where C is arbitrary constant of integration. a ( x ) = c ( x ) (cid:2) c ( x ) ddx (cid:0) f c (cid:1) − f ( x ) − b ( x ) f ( x ) (cid:3) We assume that the coefficient a ( x ) of the Riccati Eq. (1) satisfies the differential condition a ( x ) = 14 c ( x ) (cid:20) c ( x ) ddx (cid:18) f c (cid:19) − f ( x ) − b ( x ) f ( x ) (cid:21) , (34)where f ( x ) ∈ C ∞ ( I ) is an arbitrary function defined on a real interval I ⊆ ℜ .By substituting Eq. (34) into the Riccati Eq. (1), the latter can be expressed as dydx = 14 c ( x ) (cid:20) c ( x ) ddx (cid:18) f c (cid:19) − f ( x ) − b ( x ) f ( x ) (cid:21) + b ( x ) y + c ( x ) y . (35)Therefore we obtain the following: Theorem 7 . If the coefficient a ( x ) of the Riccati Eq. (1) satisfies the differential condition (34), then thegeneral solution of the Riccati Eq. (35) is given by y ( x ) = e R x [ b ( φ )+ f ( φ )] dφ C − R x c ( ψ ) e R ψ [ b ( φ )+ f ( φ )] dφ dψ + f ( x )2 c ( x ) , (36)where C is an arbitrary constant of integration. b ( x ) = f ( x ) h c ( x ) ddx (cid:0) f c (cid:1) − f ( x )2 − a ( x ) c ( x ) i We assume that the coefficient b ( x ) of the Riccati Eq. (1) satisfies the differential condition b ( x ) = 1 f ( x ) (cid:20) c ( x ) ddx (cid:18) f c (cid:19) − f ( x )2 − a ( x ) c ( x ) (cid:21) . (37)By substituting Eq. (37) into the Riccati Eq. (1), the latter can be expressed as dydx = a ( x ) + 1 f ( x ) (cid:20) c ( x ) ddx (cid:18) f c (cid:19) − f ( x )2 − a ( x ) c ( x ) (cid:21) y + c ( x ) y . (38)Therefore we obtain the following: Theorem 8 . If the coefficient b ( x ) of the Riccati Eq. (1) satisfies the differential condition (37), then thegeneral solution of the Riccati Eq. (38) is given by y ( x ) = e R x (cid:26) f φ ) (cid:20) c ( φ ) ddφ ( f c ) − f
24 ( φ )2 − a ( φ ) c ( φ ) (cid:21) + f ( φ ) (cid:27) dφ C − R x c ( ψ ) e R ψ (cid:26) f φ ) (cid:20) c ( φ ) ddφ ( f c ) − f
24 ( φ )2 − a ( φ ) c ( φ ) (cid:21) + f ( φ ) (cid:27) dφ dψ + f ( x )2 c ( x ) , (39)where C is an arbitrary constant of integration. 6 .9 Case 9: c ( x ) = f ( x ) e − R x (cid:20) f φ )2 + b ( φ ) (cid:21) dφ C +2 R x a ( ψ ) e − R ψ (cid:20) f φ )2 + b ( φ ) (cid:21) dφ dψ We assume that the coefficient c ( x ) of the Riccati Eq. (1) satisfies the integral condition c ( x ) = f ( x ) e − R x h f φ )2 + b ( φ ) i dφ C + 2 R x a ( ψ ) e − R ψ h f φ )2 + b ( φ ) i dφ dψ , (40)where C is an arbitrary constant. By substituting Eq. (40) into the Riccati Eq. (1), the latter can be expressedas dydx = a ( x ) + b ( x ) y + f ( x ) e − R x h f φ )2 + b ( φ ) i dφ C + 2 R x a ( ψ ) e − R ψ h f φ )2 + b ( φ ) i dφ dψ y . (41)Therefore we obtain the following: Theorem 9 . If the coefficient c ( x ) of the Riccati Eq. (1) satisfies the integral condition (40), then the generalsolution of the Riccati Eq. (41) is given by y ( x ) = e R x [ f ( φ )+ b ( φ )] dφ C − R x ( f ( ψ ) e − R ψ (cid:20) f φ )2 + b ( φ ) (cid:21) dφ C +2 R ψ a ( ω ) e − R ω (cid:20) f φ )2 + b ( φ ) (cid:21) dφ dω ) e R ψ [ f ( φ )+ b ( φ )] dφ dψ + 12 (cid:26) C + 2 Z x a ( ω ) e − R ω h f φ )2 + b ( φ ) i dφ dω (cid:27) e R x h f φ )2 + b ( φ ) i dφ , (42)where C is an arbitrary constant of integration. a ( x ) = b ( x ) − c ( x ) f ( x )4 c ( x ) + ddx (cid:0) − b c ± f (cid:1) We assume that the coefficient a ( x ) of the Riccati Eq. (1) satisfies the differential condition a ( x ) = b ( x ) − c ( x ) f ( x )4 c ( x ) + ddx (cid:18) − b c ± f (cid:19) , (43)where we have introduced an arbitrary function f ( x ) ∈ C ∞ ( I ) defined on a real interval I ⊆ ℜ . By substitutingEq. (43) into the Riccati Eq. (1), the latter can be expressed as dy ± dx = (cid:20) b ( x ) − c ( x ) f ( x )4 c ( x ) + ddx (cid:18) − b c ± f (cid:19)(cid:21) + b ( x ) y ± + c ( x ) y ± . (44)Therefore we obtain the following: Theorem 10 . If the coefficient a ( x ) of the Riccati Eq. (1) satisfies the differential condition (43), then thegeneral solutions of the Riccati Eq. (44) are given by y ± ( x ) = e ± R x c ( φ ) f ( φ ) dφ C ± − R x c ( ψ ) e ± R ψ c ( φ ) f ( φ ) dφ dψ − b ( x )2 c ( x ) ± f ( x ) , (45)where C ± are arbitrary constants of integration. In standard coordinates x i = ( t, r, χ, φ ) the line element for a static spherically symmetric space-time takes theform [23] ds = A ( r ) dt − V − ( r ) dr − r (cid:0) dχ + sin χdφ (cid:1) . (46)For the metric (46), Einstein’s gravitational field equations (where natural units 8 πG = c = 1 have been usedthroughout) describing the evolution of the star take the form [23] ρ ( x ) = 1 − V ( x ) x − dVdx , (47) p r ( x ) = 4 V ( x ) d ln Adx + V ( x ) − x , (48)and [1 − xη ( x )] d Adx − (cid:20) x dηdx + η ( x ) (cid:21) dAdx − (cid:20) dηdx + ∆ ( x )4 x (cid:21) A = 0 , (49)7here we have used the coordinate transformation x = r , ρ ( x ) is the energy density, p r ( x ) is the pressure in thedirection χ i (normal pressure), and p ⊥ ( x ) is the pressure orthogonal to χ i (transversal pressure), respectively. χ i is the unit space-like vector defined as χ i = √ V δ i . We assume that p r ( x ) = p ⊥ ( x ).The anisotropy parameter ∆ ( x ) is defined as ∆ ( x ) = p ⊥ ( x ) − p r ( x ). We have also defined the followingfunctions, V ( x ) = 1 − xη ( x ) , η ( r ) = m ( r ) r , (50)and 2 m ( r ) = Z r ξ ρ ( ξ ) dξ, (51)respectively. The function m ( r ) represents the mass distribution within radius r . By introducing a new function u ( x ), defined as u ( x ) = 1 A dAdx , (52)provides the metric function A ( x ) in the form A ( x ) = A e R x u ( φ ) dφ , (53)with A an arbitrary constant of integration. By substituting Eq. (52) into Eq. (49), the latter becomes a Riccatiequation of the form dudx = dηdx + ∆( x )4 x − xη ( x ) + x dηdx + η ( x )1 − xη ( x ) u − u . (54)The physical quantities η ( x ) and ∆ ( x ) are free arbitrary functions here. Now, by comparing the Riccati Eq. (54)with the Riccati Eq. (1), one gets the following settings y → u, a ( x ) → dηdx + ∆( x )4 x − xη , b ( x ) → x dηdx + η − xη , c ( x ) → − , (55)respectively. By inserting these settings into the ten theorems presented in Section II, we can obtain ten solutionsof the Riccati Eq. (54), and, consequently, ten solutions of the interior Einstein’s gravitational field equations. Inorder not to have repetitive calculations, we shall not present these results in the present paper. In the present paper, we have obtained a number of integrability conditions of the Riccati Eq. (1). All thetheorems presented in this paper have been obtained with the help of Eq. (2). By using Eq. (3), Eq. (2) can beexpressed equivalently as y [ a ( x ) , c ( x ) , y p ( x )] = y p ( x ) e R x h c ( φ ) y p ( φ ) − a ( φ ) yp ( φ ) i dφ C − R x c ( ψ ) y p ( ψ ) e R ψ h c ( φ ) y p ( φ ) − a ( φ ) yp ( φ ) i dφ dψ + y p ( x ) , (56)and y [ a ( x ) , b ( x ) , y p ( x )] = y p ( x ) e − R x h b ( φ )+ a ( φ ) yp ( φ ) i dφ C − R x h dy p ( ψ ) dψ − a ( ψ ) − b ( ψ ) y p ( ψ ) i e − R ψ h b ( φ )+ a ( φ ) yp ( φ ) i dφ dψ + y p ( x ) . (57)Note that the complexity of the general solutions Eqs. (2), (56), and (57) depends on the coefficients b ( x ) and c ( x ), a ( x ) and c ( x ), a ( x ) and b ( x ) respectively, and on the particular solution y p ( x ), satisfying the Riccati Eq. (1).Therefore, one may choose some simple particular solutions to obtain the theorems in the present paper.We have also discussed a physical application of the presented results. The study of general relativistic compactobjects is of fundamental importance for astrophysics. After the discovery of the pulsars and the explanation oftheir properties by assuming that they are rapidly rotating neutron stars, the theoretical investigation of superdensestars has been done using both numerical and analytical methods and the parameters of neutron stars have beenworked out by using a general relativistic approach.Since the Einstein’s gravitational field equations describing the interior of compact stellar objects can be reducedto a Riccati equation, the mathematical results obtained in the present paper can be used to model massive stars,such as neutron stars and pulsars. The results may be useful whenever there is some uncertainty regarding theactual equation of state of the dense matter inside the general relativistic star. Note, however, that in order tohave physical stellar models, the interior solution for static fluid spheres of Einstein’s gravitational field equationsmust satisfy specific general physical requirements. The following conditions have been generally recognized to becrucial for anisotropic fluid spheres: (i) the density ρ , the radial pressure p r and the tangential pressure p ⊥ shouldbe positive inside the star; (ii) the gradients dρ/dr , dp r /dr and dp ⊥ /dr should be negative; (iii) inside the static8onfiguration the speed of sound should be less than the speed of light, i.e. 0 ≤ dp r /dρ ≤
1, 0 ≤ dp ⊥ /dρ ≤ p r must vanish at the boundary r = R of the sphere, but the tangential pressure p ⊥ may not vanish. Thephysical interpretation of the structure and properties of the compact relativistic star models obtained by usingthe previous theorems as applied to the Riccati Eq. (54) will be presented in a future publication. We would like to thank to the anonymous referee for comments and suggestions that helped us to improve ourmanuscript. FSNL acknowledges financial support of the Funda¸c˜ao para a Ciˆencia e Tecnologia through the grantsCERN/FP/123615/2011 and CERN/FP/123618/2011.
References [1] E. Kamke,
Differentialgleichungen: L¨osungsmethoden und L¨osungen , Chelsea, New York (1959).[2] A. D. Polyanin, V. F. Zaitsev,
Handbook of exact solutions for ordinary differential equations , Boca Raton,Chapman & Hall/CRC (2003).[3] M. V. Soare, P. P. Teodorescu and I. Toma,
Ordinary differential equations with applications to mechanics ,Dordrecht, Springer (2007).[4] V. M. Strelchenya,
A new case of integrability of the general Riccati equation and its application to relaxationproblems , J. Phys. A: Math. Gen. 24, 4965 (1991).[5] E. D. Rainville,
Necessary conditions for polynomial solutions of certain Riccati equations , The AmericanMathematical Monthly , 473-476 (1936).[6] M. K. Mak, T. Harko, New integrability case for the Riccati equation , Applied Mathematics and Computation , 10974 (2012).[7] M. K. Mak, T. Harko,
New further integrability case for the Riccati equation , Applied Mathematics andComputation , 7465 (2013).[8] W. T. Reid,
Riccati Differential Equations , New York, USA: Academic Press, (1972).[9] M. K. Mak, T. Harko,
Isotropic stars in general relativity , The European Physical Journal C , 2585 (2013).[10] T. Harko, F. S. N. Lobo, M. K. Mak, Arbitrary scalar field and quintessence cosmological models ,arXiv:1310.7167 (2013).[11] T. Harko, M. K. Mak,
Anisotropic charged fluid spheres in D space-time dimensions , J. Math. Phys. , 4752(2000).[12] M. K. Mak, J. A. Belinchon, T Harko, Causal bulk viscous dissipative isotropic cosmologies with variablegravitational and cosmological constants,
International Journal of Modern Physics D, , 1265 (2002).[13] Chiang-Mei Chen, T. Harko, M. K. Mak, Viscous dissipative effects in isotropic brane cosmology , Phys. Rev.
D64 , 124017 (2001).[14] M. K. Mak, T. Harko,
Exact dissipative cosmologies with stiff fluid , Europhysics Letter. , 762 (2001).[15] Changqing Yang, Jianhua Hou, Beibo Qin, Numerical solution of Riccati differential equations by using hybridfunctions and tau method , International Journal of Mathematical and Computational Sciences , 216 (2012).[16] M. A. EI-Tawil, A. A. Bahnasawi, A. Abdel-Naby, Solving Riccati differential equation using Adomians de-composition method , Applied Mathematics and Computation , 503 (2004).[17] P. Y. Tsai, C. K. Chen,
An approximate analytical solution of the nonlinear Riccati differential equation , J.Franklin Inst. , 1850 (2011).[18] S. Abbasbandy,
Homopoty perturbation method for quadratic Riccati differential equation and comparison withAdomian decomposition method , Applied Mathematics and Computation , 485 (2006).[19] S. Abbasbandy,
Iterated He’s homopoty perturbation method for quadratic Riccati equation , Applied Mathe-matics and Computation , 581 (2006).[20] S. Abbasbandy,
A new application of He’s variational iteration method for quadratic Riccati differential equa-tion by using Adomian’s polynomials , J. Comput. Appl. Math. , 59 (2007).921] M. G¨ulsu, M. Sezer,
On the solution of the Riccati equation by the Taylor matrix method , Applied Mathematicsand Computation , 414 (2006).[22] F. Mohammadi, M. M. Hosseini,
A comparative study of numerical methods for quadratic Riccati differentialequations , J. Franklin Inst. , 156 (2011).[23] M. K. Mak, T. Harko,
Anisotropic stars in general relativity , Proc. Roy. Soc. Lond.