View
2
Download
0
Category
Preview:
Citation preview
Tilburg University
Strongly regular graphs induced by polarities of symmetric designs
Haemers, W.H.; Higman, D.G.; Hobart, S.A.
Publication date:1990
Link to publication
Citation for published version (APA):Haemers, W. H., Higman, D. G., & Hobart, S. A. (1990). Strongly regular graphs induced by polarities ofsymmetric designs. (Research memorandum / Tilburg University, Department of Economics; Vol. FEW 452).Unknown Publisher.
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
- Users may download and print one copy of any publication from the public portal for the purpose of private study or research - You may not further distribute the material or use it for any profit-making activity or commercial gain - You may freely distribute the URL identifying the publication in the public portal
Take down policyIf you believe that this document breaches copyright, please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Download date: 30. Jan. 2020
tih~ J~~ ~~0~~~o~o~o~~`~~~~~~
Q~S h~0 ~~~,IIIII I III II I IIII I IIIII II I nll ! II II I III II IIII I II
STRONGLY REGULAR GRAPHS INDUCED BYPOLARITIES OF SYNIly1ETRIC DESIGNS
W.H. HaemersD.G. Higman, S.A. Hobart
FEW 452
STRONGLY REGULAR GRAPHS INDUCED BY POLARITIES OF SYMMETRIC DESIGNS
by
W.H. HAEMERS,Tilburg University, Tilburg, The Netherlands,
D.G. HIGMAN,University of Michigan, Ann Arbor, Michigan, USA,
and
S.A. HOBART,University of Wyoming, Laramie, Wyoming, USA.
ABSTRACT. We study symmetric designs with a polarity for which the graphinduced by the absolute points and the graph induced by the non-absolutepoints both are strongly regular. The main result ís that all parameterscan be expressed in terms of only one parameter.
z
INTRODUCTION.
Throughout, we shall assume that A is the (0,1) incidence matrix of asymmetric 2-(v,k,a) design D with a polarity with v(~ 0) absolute pointsiand v- v-v (~ 0) nonabsolute points. So A takes the form:z i
A -r A C 1
Ill~ JCT A
2
where A1 and Az are symmetric matrices, A has ones on the diagonal and AZrhas zeros on the diagonal. Thus A-I and A are adjacency matrices of~ 1graphs I' and i' , say. We call i and i" the graphs induced by the polarityof D. 1'he above decomposition of A is calle~d the polar deeomposttton . Apolar decomposition is regular if I' and I'Z are regular, and strongly
reguZar if I'1 and I"2 are strongly regular, complete or empty. The incidencestructure of absolute points and nonabsolute blocks, given by the matrix C,is called the polar structure of the design D.We assume that D is nontrivial, i.e. 0~ a C k-1 ( v-2. PolardecomposiY,ions of trivi-al designs are obvious and not very interesting.
In the present paper we study strongly regular polar decompostions.It is, in a certain sense, an extension to a paper on strongly regulargraphs wit.h strongly regular decompositions by t.he First two authors [1].1'he sub,jecl. iv motivrif.ed by t.he t'ol lowinq examplc~. I,eL D be Lhe design oft,o f n i v nnrl I,yl,c~i,~, I nnc~v ol' I'(1( r~ ,~I )( u ) 2). 'I'hc~n Lho grnt,l,4 l nducc:d by Lhc~unitary polarity are strongly regular (see for instance [3]). Theparameters for this example are:
~- 4n}1-1 . k- 4n-1 I~- 4n-1-1 , V - 22n.1-(-2)n-13 3 3 ~ 3 '
We shall prove that all parameters of a strongly regular polardecomposition can be expressed in terms of only one parameter a. If a is apower of -2 we find the values above, but unfortunately, other values of afor which no examples are known, remain feasible.
3
For strongly regular I' (i - 1,2) the parameters are denoted by v, k', a1 ~ ~ ;and ui, and the eigenvalues of Ai by k, p and oi, such that ~Pí~ ) ~a ~.Note that k- k'tl and k- k:. The multiplicity of a is denoted by pi.
If fis complcte or c~mpty we put: g~ - v-1 (p disappeares).~ I I I
PRELIMINARIES.
The following three Lemmas are the analogues of 2.2, 2.4 and 2.5 of [1].Since the proofs are the same, we omit them here.
Lemma 1. Suppose I'1 is regular of degree ki - kl-1. Then
k v-v k1 1v-v i
C k~ .
Equality holds if and only if the polar decomposition is regular.
Note that for the polar decomposition to be regular k-a (the order of D)must be a square.
For a regular polar decomposition we define k2 to be the degree of I'z and
(1)k v-v ki ia - .v-vi
Then we easily have
(2) a- k~ tkz-k - tk~, ~al ~ 2.
So a, -a and k are the eigenvalues of A with multiplicities ~p (say), v-1-pand 1, respectively.
Lemma 2. Suppose A~ has eigenvalues kl, 61 and pl (loll C ~pl~) with
multiplicities 1, ~ol and vl -1-g~1 (1 (~1 ( v~ -1) , respectively. Let the
q
polar decomposition be regular. '1'hen A has eigenvalues k,-o ,-p , a, -a1 1 1 1
with multiplicities 1, pl, vr-l-pl, p-vr and vz-p, respectively.Also ~6 I~ ~Pr ~~ ~a~ . and ~o ~~ ~a~ .
Lemma 3. With the hypotheses of Lemma 2, the polar decomposition isstrongly regular if and only if one of the following occurs.i. v - v - p,
r zii. p- a, v- p,i iiii. pi - -a, vz - p,
It Ia en5,y to 5er: l.hnt tf f1 occur4 for m m~trix A, Lhen 1ti occurs for thccomplementary case, that is for the design with incidence matrix ( J denotesthe all-one matrix).
J -l AZ CTJ ~C A1
Note that Lemma 3 also applies if I'1 is complete or empty (the conditionsp- ta are meaningless ín this case). Next we show that case i does noti -occur.
Lemma 4. A strongly regular polar decomposition with vl - vz - p does notexist.
Proof. Assume v~ 2k (otherwise consider the complement). Trace A- vl -~p- k t~pa - ( v-p-1)a implies k- vr-a. Hence, by (1) and (2), kl - kz - v1~2- vz~2. Therefore, since i'z is strongly regular, kz(kz-~z-1) - uz(vz-kz-1)- u(k -1). This implies N. - 0 mod k, hence I" is complete bipartite, soz z ~ z zp, --k . This i s ímpossible , since then, by Lemma 2, ~a~ ~ ~P I- k -z z - 2 zvz~2 - v~4, which is a contradiction to aZ ( k C v~2 and ~a~ ~ 2. Q
5
MAIN TREOREM.
Theorem 5. If D has a strongly regular polar decomposition thena- 4 mod 6, and, up to taking complements, all other parameters areexpressible in terms of a as follows:
v - 3(16a~-1) , k - 3(4az-1) , ~ - 3(az-1) ,
Pi
~~ - ~ - 3(2ati)(4g-i) , ~Z - 3a(4a-1) .
k~ - k-1 - 3(a-1)(af2) , k1 - kz - 3a(2a-1) ,
-pz - a c~ - -o - az i 2 '
P~ - f~2 - 9(a-1)(2a}1) .
ul - b(a-1)(at2) , at - b(a2t4a-2) .
~z - ba(at2) , a2 - ba(a-1).
Proof. We may restrict ourselves to case ii of Lemma 3. By this result wecan express all parameters in terms of k and a only. We have a- k-az, andhence, using ~(v-1) - k(k-1):
v - k` -a~
k-a`
Trace A- v - p- k.~oa -(v-1-~p)a yieldsi
k(kta)(a-1) a(kta)(ktl-2a)(k-az)(2a-1) ~ z (k-a2)(2a-1)
Next, using (1) and (2), we find
ak k(atl)(a-1)`kz - 2a-1 ~ vZ-kz-1 -
(k-a~)(2a-1).
6
Since k-a~ - a) 0 and ~a~ ) 2 it follows that 0 C k C v-1, so I" is not- ~ z zthe empty graph or the complete graph. Moreover, 2a-1 divides k. Definex- k~(2a-1), then k- ax and a and x have the same sign. By use ofzpz - -a (by Lemma 2 and 3) and (k -p )(k -o ) - (k tp d )v (because I' isz z z z z z z z zstrongly regular) we find after some computation:
6
Define
a(axta-2x)z - 2ax-x-1 '
d - 4a -2af3 - ~a2tza-3x-3z 2ax-x-1 '
then d) 0, since d C-1 would imply (2e-1)(afl) C x(1-a) C 0, whichcontradics ~a~ ) 2. Moreover, using 2ax-x-1 - k-1-) az, we have
d C~} a 2- xtl C 4 ~ á C 5 if a and x are positive,- a 2ax-x-1 - ~-
4 2a-1 C 5 if a and x are negative.
Since d is odd we can conclude that d- 1 or d- 3. If d- 1, then x- 2a-1and ~1 -(za-i)(4a`-3at1)~(3a-1), which is no integer. Hence d- 3, so
x - 2af 1 a3 ' 6z - 2 , a-4mod6.
All other parameter now follow in a straightforward manner. 0
If a-(-2)n-1 we find the parameters of the example given in theintroduction. No other examples are known. This leaves a- 10 (D is a2-(533,133.33) design) as the smallest unsolved case. By Theorem 5, onlyfor a--2 one of the graphs induced by the polarity is complete or empty.For a--2, however, the structure is unique. So YG(2,~1) with its unitarypolarity and the complement provide the only strongly regular polardecomposition for which one of the induced graphs is empty or complete. Inthis case the polar structure is AG(2,3), the affine plane of order 3. If~a~ ) 2, it can be shown (by simular, but less complicated, arguments asfor Theorem 2.8 of [1]) that the polar structure is a strongly regulardesign (see [2] for definition and notation) with parameters
7
(3) - (v ,k-k ,a-2-X ,a-N .v ,k-k a-a ,a-u.,),i i i i ~ z' ~ ~
and t}iat, conversely, a strongly regular design whose parameters satisfy(3) with v, k, k etc. as given in Theorem 5, is the polar structure ofsome symmetric design with a strongly regular polar decomposition.
A symmetric design with a polarity can be seen as a strongly regular graphfor which loops are admitted. More precisely, i f we allow loops in graphs,a strongly regular graph is a simple strongly regular graph or a symmetricdesígn with a polarity. In this more general setting, a strongly regulargraph with strongly regular decomposition either belongs to the casewithout loops treated in [1], or is a design with a strongly regular polardecomposition. This can be proved as follows. For a strongly regulardecomposition of a symmetric design with a polarity it follows byeigenvalue arguments as in Lemma 3 , that if I'i is a symmetric design with apolarity, then so is I' . This, however is impossible by Rahilly [4]. Thuszboth I' and I' must be strongly regular graphs without loops, so we have ai 2strongly regular polar decomposition.
REFERENCES
[1] W.H. Haemers and D.G. Higman, Strongly regular graphs with stronglyregular decomposition, Linear Algebra App1. 114~115: 379-398 (1989).
[2] D.G. Hígman, Strongly regular designs and coherent configurations oftype (S .i) , European J. Combin. 9: 411-422 (1988).
[3] X.L. Hubaut, Strongly regular graphs, Discrete Math. 13: 357-381(1975).
[4] A. Rahilly, Divisions of symmeY.ric designs i nto two pnrts, Graphs andComb i nntor i cs h: fi7-73 ( l~)FiFi ).
IN 1989 REEDS VERSC}{ENEN
368 Ed Nijssen, Will Reijnders"Macht als strertegisch en tactisch mnrketinginstrr.rment, binnen dedistrihutirk~~ten"
~b~l Ra~mond GradusOptimal dynamic taxat:ion with respect to firms
370 Theo NijmanThe optimtrl choia~ of contrvls and pre-experimenLal observat.ions
371 Robert P. Gilles, Pieter }l.M. RuysRelational constraints in coalition formation
372 F.A. van der Duyn Schouten, S.G. VannesteAnalysis and computation of (n,N)-strategies for maintenance of atwo-component system
373 Drs. R. Hamers, Drs. P. VerstappenHet company ranking model: a means for evaluating the competition
374 Rommert J. CasimirInfogame Final Report
";7~i Christinn I;. MulclerI~:I'I'i~~i~~nl. nncl in~~l'I'iricnl in,:LiLul.itinrrl ni~rvrnKr~m~~nLs hr~Lwrrr~n hci-ve rnwcn Ls and L rndc un i on:; ; nn exp I arna L i un ot' h f~;h unemp I oyment ,corporatism and wlion bashing
376 Marno VerbeekOn the estimation of a fixed effects model with selective non-response
377 J. EngwerdaAdmissible target paths in economic models
378 Jack P.C. Kleijnen and Nabil AdamsPseudorandom number generation on supercomputers
379 J.P.C. BlancThe power-series algorithm applied to the shortest-queue model
380 Prof. Dr. Robert BanninkManagement's information needs and the definition of costs,with special regard to the cost of interest
381 Bert BettonvilSequential bifurcation: the design of a factor screening method
382 Bert BettonvilSequential bifurcation f'or observations with random errors
ii
383 Harold Houba and Hans KremersCorrection of the material balance equation in dynamic input-outputmodels
384 T.M. Doup, A.H. van den Elzen, A.J.J. TalmanHomotopy interpretation of price adjustment processes
385 Drs. R.T. Frambach, Prof. Dr. W.H.J. de FreytasTechnologische ontwikkeling en marketing. Een oriënterende beschou-wing
386 A.L.P.M. Hendrikx, R.M.J. Heuts, L.G. HovingGmipnriv~in r1f nul-omnt ic m~~ni Lering ~;,y5t.r~ms 1n miL~imnl ic forecnstinR
38~ Urs. J.C.I..M. WillemsEnkele opmerki.ngen over het inversíficerend gedrag van multinationaleondernemingen
388 Jack P.C. Kleijnen and Ben AnninkPseudorandom number generators revisited
389 Dr. G.W.J. HendrikseSpeltheorie en strategisch management
390 Dr. A.W.A. Boot en Dr. M.F.C.M. WijnLiquiditeit, insolventie en vermogensstructuur
391 Antoon van den Elzen, Gerard van der LaanPrice adjustment in a two-country model
392 Martin F.C.M. Wijn, Emanuel J. BijnenPrediction of failure in industryAn analysis of income statements
393 Dr. S.C.W. Eijffinger and Drs. A.P.D. GruijtersOn the short term objectives of daily intervention by the DeutscheBundesbank and the Federal Reserve System in the U.S. Dollar -Deutsche Mark exchange market
394 Dr. S.C.W. Eijffinger and Drs. A.P.D. GruijtersOn the effectiveness of daily interventions by the Deutsche Bundes-bank and the Federal Reserve System in the U.S. Dollar - DeutscheMark exchange market
395 A.E.M. Meijer and J.W.A. VingerhoetsStructural adjustment and diversification in mineral exportingdeveloping countries
396 R. GradusAbout Tobin's marginal and average qA Note
397 Jacob C. EngwerdaOn the existence-t~f a positive definite solution of the matrixequation X ~ ATX A- I
111
398 Paul C. van Batenburg and J. KriensBayesian discovery sampling: a simple model of Bayesian inference inauditing
399 Hans Kremers and Dolf TalmanSolving the nonlinear complementarity problem
400 Raymond GradusOptimal dynamic taxation, savings and investment
401 W.H. HaemersRegular two-graphs and extensions of partial geometries
402 Jack P.C. Kleijnr,n, Ren AnninkStipercompul.r,rs, Monte Carrlo simulnti~n and rregrossion ruinlysis
403 Huud 'I'. Frambach, Ed J. Nijssen, William II.J. freytasTechnologie, Strnt.egisch managemi~nt en marketing
404 Theo NijmanA natural approach to optimal forecasting in case of preliminaryobservations
405 Harry BarkemaAn empirical test of Holmstrdm's principal-agent model that tax andsignally hypotheses explicitly into account
406 Drs. W.J. van BrabandDe begrotingsvoorbereiding bij het Rijk
407 Mar~co WilkeSocietal bargaining and stability
408 Willem van Groenendaal and Aart de ZeeuwControl, coordination and conflict on international commodity markets
409 Prof. Dr. W. de Freytas, Urs. L. Arts7'ourism to Curncno: a new deal based on visitors' experiences
410 Drs. C.ti. VeLdThe use of the implied standard deviation as a predictor of futurestock price variability: a review of empirical tests
411 Drs. J.C. Caanen en Dr. E.N. KertzmanInflatieneutrale belastingheffing van ondernemingen
412 Prof. Dr. B.B. van der Genugt.enA weak law of large numbers for m-dependent rnndom variables withunbounded m
413 R.M.J. Heuts, H.P. Seidel, W.J. SelenA comparison of two lot sizing-sequencing heuristics for the processindustry
1V
414 C.B. Mulder en A.B.T.M. van SchaikEen nieuwe kijk op structuurwerkloosheid
415 Drs. Ch. CaanenDe hefboomwerking en de vermogens- en voorraadaftrek
4 1 ti Cu ido W. I mb~~nsDuration models with time-varying coefficients
41~ Guido W. ImbensEfficient estimation of choice-based sample models with the method ofmoments
418 Harry H. TigelaarOn monotone linear operators on linear spaces of square matrices
v
IN 1990 REEDS VERSCHENEN
419 Bertrand Melenberg, Rob AlessieA method to construct moments in the multi-good life cycle consump-tion model
420 J. KriensOn the differentiability of the set of efficient (N,a2) combinationsin the Markowitz portfolio selection method
421 Steffen JRlrgensen, Peter M. KortOptimal dynamic investment policies under concave-convex adjustmentcosts
422 J.P.C. BlancCyclic polling systems: limited service versus Bernoulli schedules
423 M.H.C. PaardekooperParallel normreducing transformations for the algebraic eigenvalueproblem
424 Hans GremmenOn the political (ir)relevance of classical customs union theory
42~i Ed NijtisenMarketingstrategie in Machtsperspectief
426 Jack P.C. KleijnenRegression Metamodels for Simulation with Common Random Numbers:Comparison of Techniques
42~ Harry H. TigelaarThe correlation structure of stationary bilinear processes
428 Drs. C.H. Veld en Urs. A.H.F. VerbovenDe waardering van aandelenwarrants en langlopende call-opties
429 Theo van de KlunderY, en Anton B. van SchaikLiquidity Constr,iints nnd thc Kr~ynesinn Corrfdor
430 Gert NicuwenhuisCcnt,rvr I I i m i t. t hr,nrc~m.v for sr,~~uenrr~r; w f t.h m( n )- clr,t~r~nrlr~n t. mn i n pnrt.
431 Hans J. GremmenMacro-Economic Implications of Profit Optimizing Investment Behaviour
432 J.M. SchumacherSystem-Theoretic Trends in Econometrics
433 Peter M. Kort, Paul M.J.J. van Loon, Mikulás LuptacikOptimal Dynamic Environmental Policies of a Profit Maximizing Firm
434 Raymond GradusOptimal Dynamic Profit Taxation: The Derivation of Feedback Stackel-berg Equilibria
vi
435 Jack P.C. KleijnenStatistics and Deterministic Simulation Models: Why Not?
436 M.J.G. van Eijs, R.J.M. Heuts, J.P.C. KleijnenAnalysis and comparison of two strategies for multi-item inventorysystems with joint replenishment costs
43~ Jan A. WeststrateWaiting times in a two-queue model with exhaustive and Bernoulliservice
438 Alfons DaemsTypol.ogie van non-profit organisaties
439 Drs. C.H. Veld en Drs. J. GrazellMotieven voor de uitgifte van converteerbare obligatieleningen enwarrantobligatieleningen
440 Jack P.C. KleijnenSensitivity analysis of simulation experiments: regression analysisand statistical design
441 C.H. Veld en A.H.F. VerbovenDe waardering van conversierechten van Nederlandse converteerbareobligaties
442 Drs. C.H. Veld en Drs. P.J.W. DuffhuesVerslaggevingsaspecten van aandelenwarrants
443 .Jack P.C. Kleijnen and Ren AnninkVector computers, Mcinte Carlo simulation, and regression analysis: anintroduction
411~1 Al fous: I1n~~mr;"Non-mar~kcl, fui lurus": Lmperfuctic-s in de F~udguLsccLur
445 J.P.C. BlancThe power-series algorithm applied to cyclic polling systems
446 L.W.G. Strijbosch and R.M.J. HeutsModelling (s,Q) inventory systems: parametric versus non-parametricapproximations for the lead time demand distribution
44~ Jack P.C. KleijnenSupercomputers for Monte Carlo simulation: cross-validation versusRao's test in multivariate regression
448 Jack P.C. Kleijnen, Greet van Flam and Jan RotmansTechniques for sensitivity analysis of simulation models: a casestudy of the C02 greenhouse effect
449 Harrie A.A. Verbon and Marijn J.M. VerhoevenDecision-making on pension schemes: expectation-formnCion underdemographic change
vii
450 Drs. W. Reijnders en Drs. P. VerstappenLogistiek management marketinginstrument van de jaren negentig
451 Alfons J. DaemsBudgeting the non-profit organizationAn agency theoretic approach
i iii i ~i iV i ~' ui i i ~ i ~ ii u uu ~~ i
Recommended