Internat. J. Math. & Math. Sci. VOL. 12 NO. 3 (1989) 493-502 493 A GENERALIZATION OF MULTIVALENT FUNCTIONS WITH NEGATIVE COEFFICIENTS M.K. AOUF Faculty of Science Department of Mathematics University of Mansoura Mansoura, Egypt (Received January 6, 1987) T Let ABSTRACT. the be P class of analytic and p-valent functions which can be expressed in the form n=l Tp The subclasses , Cp(A,B,a)of Tp (A,B,a) and have been considered. Sharp results concerning coefficient estimates, distortion and covering theorems are obtained. radius of convexity for the class , T (A,B,) and that the classes P C P The It is further proved T (A,B,) is determined. P (A,B,) are closed under arithmetic mean and convex linear combinations. KEY WORDS AND PHRASES. P-valent, Analytic, Radius of Convexity. 30A32, 30A36. 1980 AMS SUBJECT CLASSIFICATION CODES. I. INTRODUCTION. S (p Let f(z) I) denote the class of functions of the form P z p + ap+n n=l U there {z:Iz < I}. exists an f.. For A,B fixed, -I z p+n which are analytic and p-valent in the unit disc < F) if U, such that A function f is said to be subordinate to a function F(f analytic function (z) with l(z)l f A < B l, and 0 < p, we say that if zf’(z) f( < ] + [pB+(A-B)(p,a.)]z ’+ Bz z e U, Izl, , f e S P z (A,B,u) if and only 494 M.K. AOUF , or equivalently S f (A,B,) P zf’(z) P f(z_) B Further zf’(z) P f S --I belong to U. K (A,B,s) if class the to P T denote the subclass of P which can be expressed in the form f(z) , z . p S P P [I] have studied if Kp(A,B,a) T P [2] and Silverman and Silvia [4,5] have studied Jain certain only and we set z Certain subclasses of univalent functions with negative coefficients. Sohi and consisting of functions analytic and p-valent P T and C (A,B,a) P P (A,B,a) [3], Gupta and Silverman S ap+n n=l , (A,B,a) P I, z * (A,B,). Let T < [pB+(A-B) (p-e) f(z) said is if and only if subclasses of mmltivalent functions Also Goel and with negative In this paper we obtain coefficient estimates, distortion and covering theorems for the classes T (A,B,e) and C (A,B,e). We also determine the radius of P P It is further shown that the classes convexity for the class T (A,B,e). P T (A,B,) and C (A,B,) are closed under arithmetic mean and convex linear P P combinations. By taking =0, we get results due to Goel and Sohi [I] and by assigning specific values to A and B and taking p=l, we get results due to Silverman cbefflcients. , , , [3] and Gupta and Jain [2]. 2. COEFFICIENT INEQUALITIES. THEOREM I. A function f(z) z p [ ap+n n-I n=l [(l+B)n+(B-A)(p-e)]lap+nl.. z p+n is in T * (A,B,=) if and only if P (B-A)(p-e). (2.1) The result is sharp. zl I, then [zf’(z) -JB zf’(z) p+n -n ap+nl z n=l [pB+(A-B)(p-a)] f(z) -pf(z) " (B-A)(p-a)z p [nB + (B-A)(p-)] n=l g (l+B)n + n=l (B-A)(p-a)]]ap+nl_. Hence by the principle of maximum modulus Conversely, suppose that lap+hi z p+n] (B-A)(p-a) , f(z) P g 0. GENERALIZATION OF MULTIVALENT FUNCTIONS WITH NEGATIVE COEFFICIENTS zf" (z) P f(z) f’(z) B [pB + (A-B)(p-a)] f(z) z n--I n=l Re < (B-A) (p-=) z p- nB+(B-A)(P-)] n=l nlp+nl i n-- U. ap+n z (2.2) I. p+n zf’(z) Upon clearing the real. f(z) is through real values, we obtain z on the real axis so that denominator in (2.2) and letting E [nB+(B-A)(P-e)]Ia p+nl zp+n (B-A)(p-a)z p Choose values of 495 z (B-A)(p-=)- [nB + (B-A)(p-=)] n-- lp+nl) which implies that lap+n[ [(l+B)n + (B-A)Cp-a)] . The function z p- f(z) (B-A)Cp-a). (B-A)(p-a) (l+B)n + (B-A)(p-a) n= z p+n is an extremal function. COROLLARY I. f e T If P f(z) equality only for functions of the form A function COROLLARY 2. (B-A) (p-a) n, -(l+B)n+(BA)(p_a), ,,lap+ (A,B,=) then f(z) z p- n=l if z p with (B-A)(p-a) z (l+B)n + (B-A)(p-) lap+nlZ p+n is in p+n C (A,B,a) if and only P n-1 PROOF Since It is well known that zf’(z) P we may replace z p [ n=l ap+nl f e C (A,B,) if and only if P ap+n lp+nl z in Theorem I. zf’(z) P E T P (A,B ,a). M.K. AOUF 96 3. REPRESENTATION FORMULA. lan+plZ n+p zp f(z) A function THEOREM 2. n=l f(z) where z p exp {(B-A)(p-a) z (t) -B t@(t) P (A,B,a) if and only If (3.1) I, z e U. *(A,B,a), PROOF. zf’(z) B P f(z)-zf(z) [pB+(A-B)(p-a)]’[ f(.) < I, z Since the absolute value vanishes for z zf’(z) if(z) zf’(z) B. where h(z) o, we have (3.2) h(z) pB+( A-B (p-a) analytic is U. P f (z) h(z) T dt} l#(z)l < #(z) is analytic in U and satisfies Let f(z) e T then P is in in U and lh(z) < for z e U. Integrating (3.2) with z(z) we find that f(z) z p. exp {(B-A)(p-a) ’: (t) dt} 1-Bt#(t) The converse is obtained by differentiating (3.1). 4. DISTORTION AND COVERING THEOREMS FOR , THEOREM 3. If P rp (B-A) (p-a) I+B+(-A) (p’a) rP (B-A) (p-a) + I+B+(B-A)(p--a) with equality for PROOF. f(z) e T f(z) z , T P (A,B,a) and C P (A,B,a). (A,B,a), then rp+ f(z)l rp+ (Izl r), (B-A)(p-a) p I+B+(B-A) (p-a) (4.1) zp+l(z.+r) From Theorem I, we have n--I n-i This implies that Z ap+n n=l (B-&) (p-a) I+B+(B-A) (p-a) Thu s If(z)l Izl p + n=l>.’ lap+nl Izl p+n n=l (4.2) GENERALIZATION OF MULTIVALENT FUNCTIONS WITH NEGATIVE COEFFICIENTS r p (B-A) (p-a) + ’I-+B+(B-A) (-a) r 497 p+l Similarly, rp ) COROLLARY 3. f(z) If rp rP (B-A) (p.) p+l I+B+(B-A)(p-a) r z p Z Jan+p[ rp+ (B-A)(p-a)(p+l) p + z I+B+(B-A)(p-*)] (B-A)(p-)(p+1) n+p p (A,B,) then If(z) rP+l ([z[-- pII+B+(B-A)(p-(x)] e C r), with equality for z p- f(z) THEOREM 4. wl disc < The disc (B-A)(p-a)(p+1) z p[I+B+(B-A)(p-a)] [z[ < p+l (z +/-r). is mapped onto a domain that contains the disc , I+B I+B+(B-A)(p-a) by any [w[ < p+[pB+(A-B)(p-a)] p. [I+B+(B-A) (p-a) f e T (A,B,), and onto a domain that contains the P by any fe Cp(A B,a) The theorem is sharp, with extremal functions zp zp PROOF. (B-A) (p-a) I+B+(B-A)(p-a) (B-A)(p-a) (p+l) p[ I+B+(B-A) (p-a) zP+ Tp zp+1 THEOREM 5. If f C (A,B,a). P The results follow upon letting g (A,B,e) and r T (A,B,a), then P (P+I)(B-A),(P-a) r pr p-I- I+B+(B-A) (p-e) p [f’(z)[ in Theorem 3 and Corollary 3. M.K. AOUF 498 pr p-I (p+l) (B-A) (p-a) r p + --I+B+(B-A)(p-a) (Iz] r). Equality holds for f(z) PROOF. z (B-A)(p-) p zp+l (z I,p+nl r I+B+(B-A)(p-a) +/-r). We have If’(z) pr p-1 + Z (p+n) n--1 P rp-I + rP-I rP Z n;1 I [p+r (p+n) p+n-1 ap+nl lap+n (p+n) n-- (4.3) In view of Theorem I, Z ](I+B-)+-(A-B)=(P-a)] la n+p I+B (I+B) in+p- n;1 (B-A)(p-(x) or Z (l+B)(n+p) lan+pl p(I+B)+(A-B) (p-) . n--1 (B-A)(p-=) + Jan+p[ (4.4) (4.4) with the help of (4.2) implies that Z (n+P) an+pl !p+l) (B-A) (p-a) I+B+(B-A)(p=F A substitution of (4.5) into (4.3) yields the rlght-hand inequality. On the other-hand f’(z) ) PrP-I This completes the proof. rP-I[P -rn=ZI (p+n) (p+l) (B-A) (p-a) I+B+(B-A) (p-a) rp lap+nl (4.5) GENERALIZATION OF MULTIVALENT FUNCTIONS WITH NEGATIVE COEFFICIENTS COROLLARY 4. f(z) If z 499 p then prP-I (B-A)(p_-S_)(_+I) p-I + 5. f(z) e If Izl < R If’(z) (B-A)(p-a)(p+l)2 rP (I zl r). p+1 (z (B-A)(p-.)(p_+l) p[l+B+(B-A)(p-a)] z p RADIUS OF CONVEXITY FOR THE CLASS THEOREM 6. p p I+B+(B-A) (p-a) f(z)=z Equality holds for r +B+(B-A) (p-e) p pr 2 Tp inf n p , +/-r). , (A,B,a). (A,B,a), then f(z) is p-valently convex T P [(l+B)n+(B-A)(p-a) (B-A) (p-a) (n_p)2 in the disc n (n=l’2’ "’’)" The result is sharp, with the extremal function f(z) z (B-A) (P-a) (I+B) n + (B-A)(p-a) p zp+n We have -n z f" (z) n(n+p) an+ p zn p-n_[l (n+p)lan+plZl n n Izl n P-nl (n+p)In+pllzl n n(n+p)fan+ Thus (I + zf’(z) "- Pl p if nl (n+p)21 an+ p Izln p or nl(n-p )2 lan+pl Izl n According to Theorem 1, n.l (5.2) I" +B )n+(B-A) (p,a) (B-A)(p-a) an+ pl M.K. AOUF 500 Hence (5.2) will be true if (B-A) (p-a) or if n+-p)2]n [(l+B)n+(B-A)(p-a) (B-A) (--) Iz (5.3) (n=l’2’’’’)" The theorem follows easily from (5.3). 6. CLOSURE THEOREMS. In this section we shall prove that the classes T P (A,B,a) and C P (A,B,a) are closed under convex linear combinations. THEOREM 7. f(z) If z p n--I T P (A,B,=) h(z) then z p . n=l a n--I PROOF. Since n+ p f(z) and g(z) are in T P + b n+ p (A,B,a), n_l [(l+B)n+(B-A)(p-a)]lan+pl n+p z is also in T p (A,B,a). we have (B-A)Cp-a) (6.1) (B-A)Cp-a). (6.2) and n_l [(l+B)n+(B-A)Cp-a) ]Ibn+pl From (6.1) and (6.2) we get I/2n_ [(l+B)n+(B-A)(p-a)] h(z) which implies that T lan+p+bn+pl 6 (B-A)(p-) (A,B,a). P The following theorem can be proven similarly. n+p and g(z)fz p Cp(A,B,a), THEOREM 9. Then f Let T (A,B,a) P f(z) PROOF. zP-I/2nl lan+p+bn+p] h(z) then l_ f (z) p z p f n+p (z) z n+p is also in Cp(A.B.a). n+p (n=1,2,3,...). (B-A)(p-a) (l+B)n+(B-A)(p-a) z zp if and only if it can be expressed in the form nL kn+p fn+p (z) Suppose f(z) n=o Xn+p f where n+p (z) Xn+ p z p . o and n=o Jo Xn+p=I" (B-A) (p-a) (l+B)n+(B-A)(p-a) An+p z n+p GENERALIZATION OF MULTIVALENT FUNCTIONS WITH NEGATIVE COEFFICIENTS 501 The n Y. n So by Theorem I, n+p kn+ p I. p (A,B,a). P f(z) e T Conversely suppose Setting (I->, k f(z) e T lan+p (B-A) (p-a) (I+B)n+(B-A) (p-) IX P (A,B,). Then (B-A) (p-a) (I+B)n+(B-A)(p-) < (l+B)n+(B-A)(p-a) (BgA)(p’a) [an+p (n=l 2 ...) and P n=l n+p’ we have f(z) Z kn+p fn+p (z)" n--o This completes the proof of the theorem. (I) Putting a=o in the above theorems we get the results obtained by R.M. Goel and N.S. Sohi [I]. < I, in the above theorems we (2) Putting p=l and taking A= -B, B=, where o < get the results obtained by Gupta and Jain [2]. (3) Putting p=l and taking A= -I, B=I in the above theorems we get the results obtained by Silverman [3]. REMARKS. In conclusion, I would like to thank Professor Dr. H. Silverman for reading the manuscript and for helpful suggestions. ACKNOWLEDGMENT. REFERENCES I. GOEL, R.M. and SOHI, N.S. Multivalent Indian J. Pure appl. Math., 12(7) 2. GUPTA, V.P. and JAIN, P.K. negative coefficients Certain functions with negative coefficlents, (1981), 844-853. classes of univalent functions Bull. Austr. Math. Soc., 14 (1976), 409-416. with 502 3. M.K. AOUF SILVERMAN, H. Univalent functions with negative coefficients, Proc. Amer. Math. Soc. 51 (1975), I09-I16. 4. 5. SILVERMAN, H. and SILVIA, E.M. Convex families of starlike functions, Houston J. Math., 4, (1978), 263-268. SILVERMAN, H. and SILVIA, E.M. Pre-Starlike functions with negative coefficients Int. J. Math. Scl., 2 (1979), 427-439.