Some additive relations in the Pascal triangle

We derive some, seemingly new, curious additive relations in the Pascal triangle. They arise in summing up the numbers in the triangle along some vertical line up to some place.

Authors: A. V. Stoyanovsky

SOME ADDITIVE RELA TIONS IN THE P ASCAL TRIANGLE A. V. STOY ANOVSKY Abstract. W e derive some, see mingly new, curious additive rela- tions in the Pascal triangle. They arise in summing up the num ber s in the triangle along some vertical line up to some place. There are a lot of kno wn additive relations among ( q -)binomial co ef- ficien ts, se e, for example, [1]. In this pap er w e deriv e a see mingly new curious relatio n obtained in a simplest wa y , namely , by summation of the n um b ers of t he P ascal tria ngle along a v ertical line up to some place. It is w ell known that the sum of the n um b ers of the P ascal t r ia ngle, 1 1 1 1 2 1 1 3 3 1 1 4 6 4 1 1 5 10 10 5 1 . . . along the n - th horizon tal line equals 2 n (see, for ex ample, [2]), (1)  n 0  +  n 1  +  n 2  + . . . +  n n  = 2 n . It is also kno wn that the sum along the k -t h diagonal up to some place equals the nex t n um b er in the next diagonal, (2)  k k  +  k + 1 k  +  k + 2 k  + . . . +  n k  =  n + 1 k + 1  . Another w ell kno wn fact is that the sum along the n -th diago nal with the slope 1 / 3 equals the n -th Fib onacci n um b er, (3)  n 0  +  n − 1 1  +  n − 2 2  +  n − 3 3  + . . . = u n , where (4) u n +2 = u n +1 + u n , u 0 = u 1 = 1 1 2 A. V. S TOY ANOVSKY are the Fib onacci n um b ers. The main purp ose o f this note is t o deriv e a form ula for the sum along a ve rtical line up to some place, (5)  n k  +  n − 2 k − 1  +  n − 4 k − 2  + . . . . This form ula is the f o llo wing. Theorem. The sum (5) equals the alternated sum along the next diagonal with the slope 1 / 3, starting f rom the closest num b er, plus p ossibly ± 1 dep ending on the vertical: (6)  n k  +  n − 2 k − 1  +  n − 4 k − 2  + . . . =  n + 1 k + 1  −  n k + 2  +  n − 1 k + 3  − . . . +        0 , n − 2 k ≤ 0 , 0 , n − 2 k > 0 , n − 2 k = 6 p, 6 p + 3 , − 1 , n − 2 k > 0 , n − 2 k = 6 p + 1 , 6 p + 2 , 1 , n − 2 k > 0 , n − 2 k = 6 p − 1 , 6 p − 2 . This theorem can b e prov ed without big difficulties if one notes that b oth sides of equalit y (6 ) satisfy the same recurrence relatio n a s the n umbers in the P ascal triangle, (7) LH S ( n, k ) = LH S ( n − 1 , k ) + LH S ( n − 1 , k − 1) , RH S ( n, k ) = RH S ( n − 1 , k ) + R H S ( n − 1 , k − 1) , and the same initia l conditions (for k = 0 and k = n ). More pr ecisely , equalities (7) hold ev erywhere outside the v ertical line n − 2 k = 0 . On this line one should add 1 to the righ t hand sides of b oth equalities. Reference s [1] G. E. Andrews, T he theory of partitions, E ncyclop edia o f Mathematics and Its Applications, vol. 2, Addison-W esley , Reading , Mass.; reissue d b y Ca m bridge Univ ersity Pres s, Cambridge, 1985 . [2] I. S. Sominskii, Elementary a lgebra: a supplementary course, Fizmatlit, Moscow, 196 3 (in Russ ian). E-mail addr ess : stoyan@mcc me.ru

Original Paper

Loading high-quality paper...

Comments & Academic Discussion

Loading comments...

Leave a Comment