59 Kragujevac J. Math. 31 (2008) 59 68. BOUNDS FOR THE DISTANCE ENERGY OF A GRAPH Harishchadra S. Ramae 1, Deepak S. Revakar 1, Iva Gutma 2, Siddai Bhaskara Rao 3, B. Devadas Acharya 4 ad Haumappa B. Walikar 5 1 Departmet of Mathematics, Gogte Istitute of Techology, Udyambag, Belgaum 590008, Idia (e-mails: hsramae@yahoo.com, revakards@rediffmail.com) 2 Faculty of Sciece, Uiversity of Kragujevac, P.O. Box 60, 34000 Kragujevac, Serbia (e-mail: gutma@kg.ac.yu) 3 Stat-Math Divisio, Idia Statistical Istitute, 203, Barrackpore Road, Kolkata 700108, Idia (e-mail: raosb@isical.ac.i) 4 Departmet of Sciece ad Techology, Govermet of Idia, Techology Bhava, New Mehrauli Road, New Delhi 110016, Idia (e-mail: acharyad@alpha.ic.i) 5 Departmet of Computer Sciece, Karatak Uiversity, Dharwad 580003, Idia (e-mail: walikarhb@yahoo.co.i) (Received September 17, 2007) Abstract. The distace eergy of a graph G is defied as the sum of the absolute values of the eigevalues of the distace matrix of G. Recetly bouds for the distace eergy of
60 a graph of diameter 2 were determied. I this paper we obtai bouds for the distace eergy of ay coected graph G, thus geeralizig the earlier results. 1. INTRODUCTION I this paper we are cocered with simple graphs, that is graphs without loops, multiple or directed edges. Let G be such a graph, possessig vertices ad m edges. We say that G is a (, m)-graph. Let the graph G be coected ad let its vertices be labelled as v 1, v 2,..., v. The distace matrix of a graph G is defied as a square matrix D = D(G) = [d ij ], where d ij is the distace betwee the vertices v i ad v j i G [3,5]. The eigevalues of the distace matrix D(G) are deoted by µ 1, µ 2,..., µ ad are said to be the D-eigevalues of G. Sice the distace matrix is symmetric, its eigevalues are real ad ca be ordered as µ 1 µ 2 µ. The characteristic polyomial ad eigevalues of the distace matrix of a graph are cosidered i [6 8,14,15,18,36]. The distace eergy E D = E D (G) of a graph G is defied as [18] E D = E D (G) = µ i. The distace eergy is defied i aalogy to the graph eergy [9] E = E(G) = λ i where λ 1, λ 2,..., λ are the eigevalues of the adjacecy matrix A(G) of a graph G [5]. For more results o E(G) see [1,2,4,10 13,16,17,19 25,27 35,37]. If G = K, the complete graph o vertices, the A(K ) = D(K ) ad hece E D (G) = E(G) = 2( 1). I a recet paper [18] Idulal, Gutma, ad Vijaykumar reported lower ad upper bouds for the distace eergy of graphs whose diameter (= maximal distace betwee vertices) does ot exceed two. I this paper we obtai bouds for the distace
61 eergy of arbitrary coected (, m)-graphs, which geeralize the results obtaied i [18]. We first eed the followig Lemma. Lemma 1. Let G be a coected (, m)-graph, ad let µ 1, µ 2,..., µ be its D- eigevalues. The µ i = 0 ad µ 2 i = 2 (1) Proof. µ i = trace[d(g)] = d ii = 0. For i = 1, 2,...,, the (i, i)-etry of [D(G)] 2 is equal to d ij d ji = Hece j=1 j=1 µ 2 i = trace[d(g)] 2 = (d ij ) 2 = 2 j=1 Corollary 1.1 [18]. Let G be a coected (, m)-graph, ad let diam(g) 2, where diam(g) deotes the diameter of a graph G. The µ 2 i = 2[2 2 2 3m]. 2. BOUNDS FOR THE DISTANCE ENERGY Theorem 2. If G is a coected (, m)-graph, the 2 (d ij ) 2 E D (G) 2
62 Proof. Cosider the Cauchy Schwartz iequality ( ) 2 ( ) ( ) a i b i a 2 i b 2 i. Choosig a i = 1 ad b i = µ i, we get ( 2 µ i ) µ 2 i from which E D (G) 2 2 This leads to the upper boud for E D (G). Now ( E D (G) 2 2 = µ i ) µ i 2 = 2 (d ij ) 2 which straightforwardly leads to the lower boud for E D (G). Corollary 2.1. If G is a coected (, m)-graph, the E D (G) ( 1). Proof. Sice d ij 1 for i j ad there are ( 1)/2 pairs of vertices i G, from the lower boud of Theorem 2, E D (G) 2 (d ij ) 2 2 ( 1) 2 = ( 1). Theorem 3. Let G be a coected (, m)-graph ad let be the absolute value of the determiat of the distace matrix D(G). The 2 (d ij ) 2 + ( 1) 2/ E D (G) 2 Proof. I view of Theorem 2, we oly eed to demostrate the validity of the lower boud. This is doe aalogously to the way i which a lower boud for graph eergy is deduced i [26].
63 By defiitio of distace eergy ad Eq. (1) E D (G) 2 = = 2 ( 2 µ i ) = µ 2 i + 2 µ i µ j i i<j (d ij ) 2 + 2 i i<j µ i µ j = 2 ij ) (d 2 + µ i µ j. (2) i j Sice for oegative umbers the arithmetic mea is ot smaller tha the geometric mea, 1 ( 1) µ i µ j i j µ i µ j i j 1/( 1) ( ) 1/( 1) = µ i 2( 1) = µ i 2/ = 2/. (3) Combiig Eqs. (2) ad (3) we arrive at the lower boud. Usig Eq. (1), Corollary 1.1 ad Theorem 3 we have followig result. Corollary 3.1 [18]. Let G be a coected (, m)-graph with diam(g) 2. The 4( 1) 6m + ( 1) /2 E D (G) 2(2 2 2 3m). For a -vertex tree T [3,6], det D(T ) = ( 1) 1 ( 1)2 2 from which we obtai the followig: Corollary 3.2. For a -vertex tree T, 2 (d ij ) 2 + [( 1) +2 4 2 ] 1/ E D (T ) 2
64 Theorem 4. If G is a coected (, m)-graph, the E D (G) 2 2 (d ij ) 2 + ( 1) 2 (d ij ) 2 2 (d ij ) 2. (4) Proof. Our proof follows the ideas of Koole ad Moulto [22,23], who obtaied a aalogous upper boud for ordiary graph eergy E(G). By applyig the Cauchy Schwartz iequality to the two ( 1) vectors (1, 1,..., 1) ad ( µ 2, µ 3,..., µ ), we get 2 ( ( ) µ i ) ( 1) µ 2 i i=2 i=2 (E D (G) µ 1 ) 2 ( 1) 2 (d ij ) 2 µ 2 1 Defie the fuctio E D (G) µ 1 + ( 1) 2 f(x) = x + ( 1) 2 We set µ 1 = x ad bear i mid that µ 1 1. From µ 2 i = 2 (d ij ) 2 (d ij ) 2 µ 2 1. (d ij ) 2 x 2. we get x 2 = µ 2 1 2 (d ij ) 2 i. e. x 2 Now, f (x) = 0 implies x = 2
65 Therefore f(x) is a decreasig fuctio i the iterval 2 (d ij ) 2 x 2 (d ij ) 2 ad 2 Hece ad iequality (4) follows. (d ij ) 2 2 f(µ 1 ) f 2 (d ij ) 2 µ 1. (d ij ) 2 From Eq. (1), Corollary 1.1, ad Theorem 4 we obtai: Corollary 4.1. Let G be a coected (, m)-graph with diam(g) 2. The E D (G) 42 4 6m ( ) 2 4( 1) 6m + ( 1) 4( 1) 6m. Refereces [1] R. Balakrisha, The eergy of a graph,li. Algebra Appl. 387 (2004), 287 295. [2] V. Brakov, D. Stevaović, I. Gutma, Equieergetic chemical trees, J. Serb. Chem. Soc. 69 (2004), 549 553. [3] F. Buckley, F. Harary, Distace i Graphs, Addiso Wesley, Redwood, 1990. [4] A. Che, A. Chag, W. C. Shiu, Eergy orderig of uicyclic graphs, MATCH Commu. Math. Comput. Chem. 55 (2006), 95 102. [5] D. M. Cvetković, M. Doob, H. Sachs, Spectra of Graphs Theory ad Applicatios, Academic Press, New York, 1980.
66 [6] M. Edelberg, M. R. Garey, R. L. Graham, O the distace matrix of a tree, Discrete Math. 14 (1976), 23 39. [7] R. L. Graham, L. Lovász, Distace matrix polyomial of trees, Adv. Math. 29 (1978), 60 88. [8] R. L. Graham, H. O. Pollak, O the addressig problem for loop switchig, Bell System Tech. J. 50 (1971), 2495 2519. [9] I. Gutma, The eergy of a graph, Ber. Math. Stat. Sekt. Forschugsz. Graz 103 (1978), 1 22. [10] I. Gutma, Topology ad stability of cojugated hydrocarbos. The depedece of total π-electro eergy o molecular topology, J. Serb. Chem. Soc. 70 (2005), 441 456. [11] I. Gutma, B. Furtula, H. Hua, Bipartite uicyclic graphs with maximal, secod maximal ad third maximal eergy, MATCH Commu. Math. Comput. Chem. 58 (2007), 85 92. [12] I. Gutma, O. E. Polasky, Mathematical Cocepts i Orgaic Chemistry, Spriger Verlag, Berli, 1986. [13] I. Gutma, S. Zare Firoozabadi, J. A. de la Peña, J. Rada, O the eergy of regular graphs, MATCH Commu. Math. Comput. Chem. 57 (2007), 435 442. [14] A. J. Hoffma, M. H. McAdrew, The polyomial of a directed graph, Proc. Amer. Math. Soc. 16 (1965), 303 309. [15] H. Hosoya, M. Murakami, M. Gotoh, Distace polyomial ad characterizatio of a graph, Natur. Sci. Rept. Ochaumizu Uiv. 24 (1973), 27 34. [16] H. Hua, O miimal eergy of uicyclic graphs with prescribed girth ad pedat vertices, MATCH Commu. Math. Comput. Chem. 57 (2007), 351 361.
67 [17] H. Hua, Bipartite uicyclic graphs with large eergy, MATCH Commu. Math. Comput. Chem. 58 (2007), 57 83. [18] G. Idulal, I. Gutma, A. Vijaykumar, O the distace eergy of a graph, MATCH Commu. Math. Comput. Chem. 60 (2008), 461 472. [19] G. Idulal, A. Vijaykumar, O a pair of equieergetic graphs, MATCH Commu. Math. Comput. Chem. 55 (2006), 83 90. [20] G. Idulal, A. Vijaykumar, Eergies of some o-regular graphs, J. Math. Chem. 42 (2007), 377 386. [21] G. Idulal, A. Vijaykumar, Equieergetic self-complemetary graphs, Czechoslovak Math. J., to appear. [22] J. Koole, V. Moulto, Maximal eergy graphs, Adv. Appl. Math. 26 (2001), 47 52. [23] J. Koole, V. Moulto, Maximal eergy bipartite graphs, Graph. Combi. 19 (2003), 131 135. [24] W. Li, X. Guo, H. Li, O the extremal eergies of trees with a give maximum degree, MATCH Commu. Math. Comput. Chem. 54 (2005), 363 378. [25] F. Li, B. Zhou, Miimal eergy of bipartite uicyclic graphs of a give bipartitio, MATCH Commu. Math. Comput. Chem. 54 (2005), 379 388. [26] B. J. McClellad, Properties of the latet roots of a matrix: The estimatio of π-electro eergies, J. Chem. Phys. 54 (1971), 640 643. [27] H. S. Ramae, H. B. Walikar, Costructio of equieergetic graphs, MATCH Commu. Math. Comput. Chem. 57 (2007), 203 210. [28] H. S. Ramae, H. B. Walikar, S. B. Rao, B. D. Acharya, P. R. Hampiholi, S.R. Jog, I. Gutma, Equieergetic graphs, Kragujevac J. Math. 26 (2004), 5 13.
68 [29] H. S. Ramae, H. B. Walikar, S. B. Rao, B. D. Acharya, P. R. Hampiholi, S.R. Jog, I. Gutma, Spectra ad eergies of iterated lie graphs of regular graphs, Appl. Math. Lett. 18 (2005), 679 682. [30] H. B. Walikar, H. S. Ramae, I. Gutma, S. B. Halkari, O equieergetic graphs ad molecular graphs, Kragujevac J. Sci. 29 (2007), 73 84. [31] W. Ya, L. Ye, O the maximal eergy of trees with a give diameter, Appl. Math. Lett. 18 (2005), 1046 1052. [32] A. Yu, M. Lu, F. Tia, New upper boud for the eergy of graphs, MATCH Commu. Math. Comput. Chem. 53 (2005), 441 448. [33] L. Ye, X. Yua, O the miimal eergies of trees with a give umber of pedet vertices, MATCH Commu. Math. Comput. Chem. 57 (2007), 193 201. [34] B. Zhou, Eergy of a graph, MATCH Commu. Math. Comput. Chem. 51 (2004), 111 118. [35] B. Zhou, Lower bouds for the eergy of quadragle free graphs, MATCH Commu. Math. Comput. Chem. 55 (2006), 91 94. [36] B. Zhou, O the largest eigevalue of the distace matrix of a tree, MATCH Commu. Math. Comput. Chem. 58 (2007), 657 662. [37] B. Zhou, I. Gutma, J. A. de la Peña, J. Rada, L. Medoza, O spectral momets ad eergy of graphs, MATCH Commu. Math. Comput. Chem. 57 (2007), 183 191.