Citation: Xiaodan Yuan, Jiagui Luo. On the Diophantine equation 1x+1y+1z=13p[J]. AIMS Mathematics, 2017, 2(1): 111-127. doi: 10.3934/Math.2017.1.111
[1] | Wei Zhao, Jian Lu, Lin Wang . On the integral solutions of the Egyptian fraction equation ap=1x+1y+1z. AIMS Mathematics, 2021, 6(5): 4930-4937. doi: 10.3934/math.2021289 |
[2] | Jinyan He, Jiagui Luo, Shuanglin Fei . On the exponential Diophantine equation (a(a−l)m2+1)x+(alm2−1)y=(am)z. AIMS Mathematics, 2022, 7(4): 7187-7198. doi: 10.3934/math.2022401 |
[3] | Bingzhou Chen, Jiagui Luo . On the Diophantine equations x2−Dy2=−1 and x2−Dy2=4. AIMS Mathematics, 2019, 4(4): 1170-1180. doi: 10.3934/math.2019.4.1170 |
[4] | Cheng Feng, Jiagui Luo . On the exponential Diophantine equation (q2l−p2k2n)x+(pkqln)y=(q2l+p2k2n)z. AIMS Mathematics, 2022, 7(5): 8609-8621. doi: 10.3934/math.2022481 |
[5] | Jing Huang, Ao Han, Huafeng Liu . On a Diophantine equation with prime variables. AIMS Mathematics, 2021, 6(9): 9602-9618. doi: 10.3934/math.2021559 |
[6] | Jing Huang, Qian Wang, Rui Zhang . On a binary Diophantine inequality involving prime numbers. AIMS Mathematics, 2024, 9(4): 8371-8385. doi: 10.3934/math.2024407 |
[7] | Changsheng Luo, Jiagui Luo . Complete solutions of the simultaneous Pell equations (a2+1)y2−x2=y2−bz2=1. AIMS Mathematics, 2021, 6(9): 9919-9938. doi: 10.3934/math.2021577 |
[8] | Xiao Jiang, Shaofang Hong . On the denseness of certain reciprocal power sums. AIMS Mathematics, 2019, 4(3): 412-419. doi: 10.3934/math.2019.3.412 |
[9] | Liuying Wu . On a Diophantine equation involving fractional powers with primes of special types. AIMS Mathematics, 2024, 9(6): 16486-16505. doi: 10.3934/math.2024799 |
[10] | Bhabesh Das, Helen K. Saikia . On the Sum of Unitary Divisors Maximum Function. AIMS Mathematics, 2017, 2(1): 96-101. doi: 10.3934/Math.2017.1.96 |
In 1950, Erdös [1] conjectured for any positive integer n>1, that the following diophantine equation
1x+1y+1z=4n | (1.1) |
has positive integer solutions x, y, z. Later, Strauss [1] made a more powerful conjecture: let n>2, then diophantine equation (1.1) has positive integer solution x,y,z with x≠y,x≠z,y≠z. He proved that the conjecture is true with n<5000. In 1964, Zhao Ke, Qi Sun and Xianjue Zhang [3] proved that Strauss conjecture is equivalent to Erdös conjecture. In 1979, Ke and Sun [4] proved that Erdös-Strauss conjecture is true with n<4⋅105. In 1965, Yamamoto [10] proved that Erdös-Strauss conjecture is also true with n<107. In 1978, Franceschine [2] proved that Erdös-Strauss conjecture is true with n<108. Sierpiski made a similar conjecture: for any positive integer n>1, that the following diophantine equation
1x+1y+1z=5n | (1.2) |
has the solutions x,y,z of positive integer. Palama [6,7] proved that Sierpiski conjecture is true with n<922321.Stewart [9] also obtained above the result with n≤105743881 and n≢1(mod278460). In 1984, Liu [5] obtained all the solution of positive integers of the following diophantine equation
1x+1y+1z=5121. | (1.3) |
Write n=12p in (1.1) or write n=15p in (1.2), where p is an odd prime. We obtain diophantine equation
1x+1y+1z=13p,x≤y≤z. | (1.4) |
One can easily get all the solutions of positive integer of (1.4) when p|x. If 3|x,p∤x, then one lets
x=3ax1,y=ay1,z=az1,a=gcd(x/3,y,z), |
and so (1.4) is changed to
3p(y1+z1)x1=(ax1−p)y1z1. | (1.5) |
Moreover, we write y1=dy2,z1=dz2,d=gcd(y1,z1). It is easy to see that gcd(d,x1)=1.
If 3p|d, then we get that x1|y2z2 and y2z2|x1. It follows that x1=y2z2. Write d3p=d1, then (1.5) is changed to
y2+z2d1=ay2z2−p,gcd(y2,z2)=1,y2≤z2. |
If 3∤d,p|d and 3|y1, then we get that x1|y3z2 and y3z2|x1 where y2=3y3. It follows that x1=y3z2. Write dp=d1, then (1.5) is changed to
3y3+z2d1=ay3z2−p,gcd(3y3,z2)=1,3y3≤z2. |
If 3∤d,p|d and 3|z1, then we get that x1|y2z3 and y2z3|x1 where z2=3z3. It follows that x1=y2z3. Write dp=d1, then (1.5) is changed to
y2+3z3d1=ay2z3−p,gcd(y2,3z3)=1. |
If p∤d,3|d and p|y1, then we get that x1|y3z2 and y3z2|x1 where y2=py3. It follows that x1=y3z2. Write d3=d1, then (1.5) is changed to
py3+z2d1=ay3z2−p,gcd(py3,z2)=1,py3≤z2. |
If p∤d,3|d and p|z1, then we get that x1|y2z3 and y2z3|x1 where z2=pz3. It follows that x1=y2z3. Write d3=d1, then (1.5) is changed to
y2+pz3d1=ay2z3−p,gcd(y2,pz3)=1. |
If 3p∤d and 3|y1,p|z1, then we get that x1|y3z3 and y3z3|x1 where y2=3y3,z2=pz3. It follows that x1=y3z3. Then (1.5) is changed to
3y3+pz3d=ay3z3−p,gcd(3y3,pz3)=1. |
If 3p∤d and p|y1,3|z1, then we get that x1|y3z3 and y3z3|x1 where y2=py3,z2=3z3. It follows that x1=y3z3. Then (1.5) is changed to
py3+3z3d=ay3z3−p,gcd(py3,3z3)=1. |
If 3∤y1z1,p∤d and p|y1, then we get that x1|y3z2 and y3z2|x1 where y2=py3. It follows that x1=y3z2. Then (1.5) is changed to
py3+z2d=ay3z2−p3,gcd(py3,z2)=1. |
If 3∤y1z1,p∤d and p|z1, then we get that x1|y2z3 and y2z3|x1 where z2=pz3. It follows that x1=y2z3. Then (1.5) is changed to
y2+pz3d=ay2z3−p3,gcd(y2,pz3)=1. |
If 3∤y1z1,p|d, then we get that x1|y2z2 and y2z2|x1. It follows that x1=y2z2. Write dp=d1, then (1.5) is changed to
y2+z2d1=ay2z2−p3,gcd(y2,z2)=1. |
Thus we have proved that solving the equation (1.5) is equivalent to solving the following diophantine equations:
y+zd=ayz−p,gcd(y,z)=1,y≤z, | (1.6) |
3y+zd=ayz−p,gcd(3y,z)=1, | (1.7) |
py+zd=ayz−p,gcd(py,z)=1, | (1.8) |
3y+pzd=ayz−p,gcd(3y,pz)=1, | (1.9) |
py+zd=ayz−p3,gcd(py,z)=1,3∤yz, | (1.10) |
y+zd=ayz−p3,gcd(y,z)=1,y≤z,3∤yz. | (1.11) |
In this paper, we investigate the equations (1.6), (1.7) and (1.8) with p=661. Actually, we get all the solutions of positive integer of them. That is, we have the following results:
Theorem 1.1. If ayz−661≥80, then all the solutions of positive integers of (1.6) are given by (a,d,y,z)=
(1, 1, 2, 663), (1, 1, 3, 332), (1, 2, 1, 1323), (1, 2, 3, 265), (1, 3, 1, 992), (1, 3, 2, 397),
(2, 1, 1, 662), (2, 1, 2, 221), (2, 1, 4, 95), (2, 2, 1, 441), (3, 1, 1, 331),
(3, 1, 3, 83), (3, 3, 1, 248)(4, 2, 1, 189).
Theorem 1.2. If ayz−661≥90, then all the solutions of positive integers of (1.7) are given by (a,d,y,z)=
(1, 1, 2, 667), (1, 1, 3, 335), (1, 1, 5, 169), (1, 1, 9, 86), (1, 1, 84, 11), (1, 1, 167, 7), (1, 1, 333, 5),
(1, 1, 665, 4), (1, 2, 1, 1325), (1, 4, 2645, 1), (1, 5, 1, 827), (1, 5, 1653, 1), (1, 7, 1157, 1), (1, 11, 909, 1),
(1, 19, 785, 1), (2, 1, 1, 664), (2, 1, 3, 134), (2, 1, 27, 14), (2, 1, 133, 4), (2, 1, 663, 2), (2, 2, 1323, 1),
(2, 4, 529, 1), (3, 1, 1, 332), (3, 1, 221, 2), (3, 2, 1, 265), (3, 2, 441, 1), (4, 1, 662, 1), (5, 1, 1, 166),
(5, 1, 331, 1), (5, 2, 189, 1), (6, 2, 147, 1), (8, 1, 51, 2), (10, 1, 39, 2), (12, 2, 63, 1).
Theorem 1.3. Each of the following is true:
1. If ayz−661<80, then all the solutions of positive integers of (1.6) are given by (a,d,y,z)=
(1, 3, 5, 142), (1, 3, 17, 40), (1, 7, 6, 113), (1, 11, 18, 37), (1, 28, 17, 39), (1, 32, 13, 51),
(1, 67, 2, 333), (1, 112, 3, 221), (1, 332, 1, 663), (1, 333, 2, 331), (1, 663, 1, 662), (2, 1, 5, 74),
(2, 1, 11, 32), (2, 1, 14, 25), (2, 2, 11, 31), (2, 7, 1, 356), (2, 8, 5, 67), (2, 20, 1, 339), (2, 26, 1, 337),
(2, 29, 4, 83), (2, 112, 1, 332), (2, 332, 1, 331), (3, 1, 11, 21), (3, 14, 13, 17), (3, 28, 1, 223),
(3, 111, 1, 221), (4, 1, 6, 29), (4, 5, 1, 174), (4, 24, 1, 167), (5, 1, 4, 35), (5, 1, 6, 23), (5, 2, 1, 147),
(5, 15, 1, 134), (6, 4, 1, 115), (6, 8, 3, 37), (7, 1, 2, 51), (7, 6, 5, 19), (7, 24, 1, 95), (8, 28, 1, 83),
(9, 1, 2, 39), (9, 1, 4, 19), (9, 2, 3, 25), (9, 15, 1, 74), (11, 2, 1, 63), (11, 3, 1, 62), (13, 10, 3, 17),
(13, 26, 1, 51), (14, 2, 1, 49), (15, 1, 5, 9), (17, 8, 3, 13), (17, 20, 1, 39), (19, 3, 5, 7), (19, 9, 1, 35),
(20, 1, 2, 17), (21, 3, 1, 32), (23, 5, 1, 29), (24, 1, 4, 7), (25, 2, 1, 27), (26, 1, 2, 13), (28, 1, 3, 8),
(29, 4, 1, 23), (32, 2, 1, 21), (35, 5, 1, 19), (39, 9, 1, 17), (51, 7, 1, 13), (74, 2, 1, 9), (83, 3, 1, 8),
(95, 2, 1, 7), (111, 1, 2, 3), (221, 2, 1, 3), (331, 3, 1, 2), (332, 1, 1, 2), (662, 2, 1, 1), (663, 1, 1, 1).
2. If ayz−661<90, then all the solutions of positive integers of (1.7) are given by (a,d,y,z)=
(1, 9, 1, 744), (1, 29, 3, 223), (1, 31, 35, 19), (1, 37, 5, 133), (1, 67, 39, 17), (1, 73, 95, 7),
(1, 83, 51, 13), (1, 101, 133, 5), (1, 115, 3, 221), (1, 127, 677, 1), (1, 167, 1, 665), (1, 333, 1, 663),
(1, 337, 2, 331), (1, 665, 1, 662), (1, 995, 331, 2), (1, 995, 663, 1), (1, 1987, 662, 1), (2, 1, 13, 28),
(2, 2, 3, 121), (2, 200, 333, 1), (2, 334, 1, 331), (2, 994, 331, 1), (3, 2, 4, 58), (3, 7, 2, 113),
(3, 11, 6, 37), (3, 28, 13, 17), (3, 32, 17, 13), (3, 67, 111, 2), (3, 112, 1, 221), (3, 332, 221, 1),
(4, 2, 9, 19), (5, 10, 7, 19), (5, 16, 19, 7), (5, 34, 1, 133), (5, 100, 133, 1), (6, 5, 16, 7), (6, 20, 113, 1),
(8, 1, 18, 5), (8, 23, 84, 1), (9, 1, 1, 83), (9, 1, 7, 11), (12, 1, 2, 29), (12, 5, 58, 1), (13, 2, 1, 53),
(13, 13, 3, 17), (15, 1, 2, 23), (15, 2, 49, 1), (17, 11, 3, 13), (17, 59, 39, 1), (18, 8, 1, 37), (26, 2, 27, 1),
(27, 1, 13, 2), (27, 2, 1, 25), (33, 2, 21, 1), (34, 1, 5, 4), (37, 11, 18, 5), (39, 10, 1, 17), (39, 26, 17, 1),
(45, 1, 3, 5), (51, 8, 1, 13), (51, 20, 13, 1), (75, 2, 9, 1), (84, 1, 1, 8), (96, 2, 7, 1), (112, 1, 3, 2),
(133, 2, 1, 5), (167, 1, 1, 4), (221, 5, 3, 1), (222, 2, 3, 1), (331, 5, 1, 2), (331, 7, 2, 1), (333, 1, 1, 2),
(334, 1, 2, 1), (662, 4, 1, 1), (663, 2, 1, 1), (665, 1, 1, 1).
Theorem 1.4. All the solutions of positive integers of (1.8) with py≤z are given by (a,d,y,z)=
(1, 1, 2, 1983), (1, 2, 1, 1983), (1, 3, 1, 1322), (1, 3, 2, 661), (1, 331, 2, 333), (1, 662, 1, 663),
(1, 1323, 1, 662), (1, 1653, 2, 331), (2, 1, 1, 1322), (2, 2, 1, 661), (2, 331, 1, 332), (2, 992, 1, 331).
We organize this paper as follows. In Section 2, we present some lemmas which are needed in the proof of our main results. Consequently, in Sections 3 to 6, we give the proofs of Theorem 1.1 to 1.4 respectively.
To prove the main theorems, we need the following lemmas.
Lemma 2.1. Let 661+b=p, where p is an odd prime. If one of the following conditions is satisfied
p≡1(modq),q=3,5or11orp≡3(mod7),orp≡23(mod29),orp≡5(mod41), |
then both equation (1.6) and equation (1.7) have no solution (a, d, y, z) of positive integers with ayz−661=b.
proof. We only prove the case p≡23(mod29). The proofs of other cases are similarly. By the assumption we have b≡p−661≡0(mod29). Assume either (1.6) or (1.7) has a positive integer solution (a,d,y,z) with ayz−661=b. It follows (y,z)∈{(1,1),(1,p),(p,1)} and either
y+z≡0(mod29) | (2.1) |
or
3y+z≡0(mod29). | (2.2) |
One can easily to see neither (2.1) nor (2.2) is satisfied for any element (y,z)∈{(1,1),(1,p),(p,1)}. This completes the proof of Lemma 2.1}.
Lemma 2.2. Let 661+b=pq, where p,q are prime numbers.
(i)If one of the following conditions is satisfied
p≡q≡1(mod3),orp≡q≡1(mod5),orp≡1(mod7),q≡3(mod7),orp≡3(mod13),q≡−5(mod13),orp≡3(mod17),q≡5(mod17),orp≡3(mod19),q≡5(mod19),orp≡3(mod20),q≡7(mod20),orp≡7(mod23),q≡9(mod23),orp≡3(mod31),q≡−7(mod31),orp≡2(mod37),q≡14(mod37),orp≡23(mod52),q≡31(mod52),orp≡3(mod56),q≡15(mod56),orp≡2(mod57),q≡17(mod57),orp≡q≡2(mod73),orp≡2(mod85),q≡33(mod85),orp≡7(mod88),q≡19(mod88), |
then both equation (1.6) and equation (1.7) have no positive integer solution (a, d, y, z) with ayz−661=b.
(ii). If one of the following conditions is satisfied
p≡q≡1(mod4),orp≡q≡±2(mod9),orp≡2(mod11),q≡6(mod11), |
then equation (1.6) has no positive integer solution (a, d, y, z) with ayz−661=b.
proof. (i) We only prove the case p≡7(mod23),q≡9(mod23).The proofs of other cases are similarly. By the assumption we have b≡pq−661≡0(mod23). Assume either (1.6) or (1.7) has a positive integer solution (a,d,y,z) with ayz−661=b. It follows (y,z)∈{(1,puqv),(p,qv),(q,pu),(pq,1),u,v=0,1} and either
y+z≡0(mod23) | (2.3) |
or
3y+z≡0(mod23). | (2.4) |
One can easily to see neither (2.3) nor (2.4) is satisfied for any element (y,z)∈{(1,puqv),(p,qv),(q,pu),(pq,1),u,v=0,1}. This completes the proof of part (i) of Lemma 2.2.
The proof of part (ii) is similar.
Lemma 2.3. Let 661+b=pqr, where p,q,r are prime. If one of the following conditions is satisfied
p≡q≡3(mod11),r≡5(mod11),orp≡q≡3(mod25),r≡4(mod25),orp≡q≡2(mod31),r≡18(mod31),orp≡q≡2(mod55),r≡14(mod55),orp≡q≡19(mod61),r≡2(mod61),orp≡q≡5(mod64),r≡29(mod64),orp≡3(mod80),q≡13(mod80),r≡19(mod80), |
then both equation (1.6) and equation (1.7) have no positive integer solution (a, d, y, z) with ayz−661=b.
proof. We only prove the case p≡q≡19(mod61),r≡2(mod61). The proofs of other cases are similarly. By the assumption we have b≡pqr−661≡0(mod61). Assume either (1.6) or (1.7) has a positive integer solution (a,d,y,z) with ayz−661=b. It follows
(y,z)∈{(1,puqvrt),(p,qvrt),(q,purt),(r,puqv),(pq,rt),(pr,qv),(qr,pt),(pqr,1),u,v,t=0,1} |
and either
y+z≡0(mod61) | (2.5) |
or
3y+z≡0(mod23). | (2.6) |
One can easily to see neither (2.5) nor (2.6) is satisfied for any element (y,z)∈{(1,puqvrt),(p,qvrt),(q,purt),(r,puqv),(pq,rt),(pr,qv),(qr,pt),(pqr,1),u,v,t=0,1}. This completes the proof of Lemma 2.3.
Lemma 2.4. If 3|b, then equation (1.7) has no positive integer solution (a, d, y, z) with ayz−661=b.
proof. Assume (1.7) has a positive integer solution (a,d,y,z) with ayz−661=b. Then we get b|3y+z. It implies that 3|z since 3|b, which contradicts with gcd(3,z)=1. This completes the proof of Lemma 2.4.
Lemma 2.5. Let 661+b=2kp, where k is an integer more than 2 and p is an odd prime.
If 1+2upv≢0(modb),p+2u≢0(modb), and 2t+pv≢0(modb), where v∈{0,1},0≤u≤k,1≤t≤k, then equation (1.6) has no positive integer solution (a, d, y, z) with ayz−661=b.
proof. Assume (1.6) has a positive integer solution (a,d,y,z) with ayz−661=b. It follows
(y,z)∈{(1,2upv),(p,2u),(2t,pv),(2tp,1),0≤u≤k,v=0,1,1≤t≤k} |
and
y+z≡0(modb), | (2.7) |
which contradicts the assumption of Lemma 2.5.
This completes the proof of Lemma 2.5.
Lemma 2.6. Let (a,d,y,z) be a positive integer solution of equation (1.6) with ayz−661=b. If b≥80,ad≥2, then y<6.
proof. We first prove that b≤1322. Otherwise b>1322, then we have z>y+z2≥13222=661.
If ay>2, then from ayz−661=y+zd≤2z, we get z≤(ay−2)z≤661 which contradicts with z>661.
If a=y=d=1, then we have 1=−661 which is impossible.
If a=y=1,d>1, then we have
z=661d+1d−1=661+662d−1≤661+662=1323, |
which contradicts with z>1322+661=1983.
If a=1,y=2, then we have
z=6612+6652(2d−1)≤661+6652=663, |
which contradicts with z>1322+661=1983.
If a=2,y=1, then we have
z=6612+6632(2d−1)≤661+6632=662, |
which also contradicts with z>1322+661=1983. Hence b≤1322 as desired.
Assume now y≥6.
If d≥2,b≥200, then y+z=db≥400. Since quadratic function
f(y)=y(db−y),0<y≤db2, |
is increasing function, so we have
yz=f(y)≥f(6)=6⋅(db−6)≥6⋅394=2364, |
which contradicts with ayz≤1983. If d≥2,80≤b<200, then similarly we have
yz=f(y)≥f(6)=6⋅(db−6)≥6⋅154=924, |
which contradicts with ayz≤661+200=861.
If d=1, then we have a≥2. It follows that yz≤[1983a]≤[19832]=991. If b≥200, then similarly we have
yz=f(y)≥f(6)=6⋅(b−6)≥6⋅194=1164, |
which contradicts with yz≤991. If 80≤b<200, then similarly we have
yz=f(y)≥f(6)=6⋅(b−6)≥6⋅74=444, |
which contradicts with yz≤[861a]≤[8612]=430.
This completes the proof of Lemma 2.6.
Lemma 2.7. Let (a, d, y, z) be a positive integer solution of equation (1.7) with ayz−661=b. If 80≤b≤2644, and a≥5 or d≥7, then we have either y<6 or z<6.
proof. Since gcd(3y,z)=1, so we have either 3y<z or 3y>z. We prove that y<6 if 3y<z and z<6 if 3y>z.
We prove that z<6 if 3y>z. Otherwise z≤6.
Case 1: d≥7. Then 3y+z=db, we get z≤(ay−2)z≤661 which contradicts with z>661.
If b≥300, then y+z=db≥2100. Since quadratic function
g(z)=z(db−z),0z≤db2, |
is increasing function, so we have
yz=g(z)3≥g(6)3≥2⋅2094=4188, |
which contradicts with ayz≤3305.
If 90≤b<300, then similarly we have
yz=g(z)3≥g(6)3≥2⋅624=1248, |
which contradicts with ayz≤661+300=961.
Case 2: a≥5, then we have yz≤[3305a]≤[33055]=661.
If b≥400, then similarly we have
yz=g(z)3≥g(6)3≥2⋅394=788, |
which contradicts with yz≤661.
If 150≤b<400, then similarly we have
yz=g(z)3≥g(6)3≥2⋅144=288, |
which contradicts with yz≤[1061a]≤[10615]=212.
If 90≤b<150, then similarly we have
yz=g(z)3≥g(6)3≥2⋅84=168, |
which contradicts with yz≤[811a]≤[8115]=162.
Similarly we can prove that y<6 if 3y<z.
This completes the proof of Lemma 2.7.
Assume that (a,d,y,z) is a positive integer solution of equation (1.6) with ayz−661=b≥80. Then we have y≤5 by Lemma :2.6. We divide the proof into four cases.
Case 1: d=1. From equation (1.6), we get
(ay−1)z=661+y. | (3.1) |
Replacing y by 1,2,3,4,5 in (3.1) respectively, we obtain
(a−1)z=2⋅331,(2a−1)z=3⋅13⋅17,(3a−1)z=8⋅83,(4a−1)z=5⋅7⋅19,(5a−1)z=2⋅9⋅37. |
Thus we get
(a,y,z)=(2,1,662),(3,1,331),(2,2,221),(3,3,83),(2,4,95). |
Case 2: d=2. From equation (1.6), we get
(2ay−1)z=1322+y. | (3.2) |
Replacing y by 1,2,3,4,5 in (3.2) respectively, we obtain
(2a−1)z=9⋅49,(4a−1)z=4⋅331,(6a−1)z=25⋅53,(8a−1)z=6⋅13⋅17,(10a−1)z=1327. |
Thus we get
(a,y,z)=(1,1,1323),(2,1,441),(4,1,189),(1,3,265).a |
Case 3: d=3. From equation (1.6), we get
(3ay−1)z=1983+y. | (3.3) |
Replacing y by 1,2,3,4,5 in (3.3) respectively, we obtain
(3a−1)z=43⋅31,(6a−1)z=5⋅397,(9a−1)z=6⋅331,(12a−1)z=1987,(15a−1)z=4⋅7⋅71. |
Thus we get
(a, y, z)=(1, 1, 992), (3, 1, 248), (1, 2, 397). |
Case 4: d\geq 4. If b\geq 100, then by the proof of Lemma 2.6 we know
yz=f(y)\ge f(6)\ge 6\cdot 394=2364, |
which contradicts with ayz\leq 1983. If 80\leq b<100, then similarly we know
yz=f(y)\ge f(6)\ge 6\cdot 314=1884, |
which contradicts with ayz\leq 761.
So Theorem 1.1 is proved.
Assume that (a, d, y, z) is a positive integer solution of equation (1.7) with ayz-661=b\geq 90. We divide the proof into three cases.
Case 1: b>2644, assume that y>z. Then y>\frac{3y+z}{4}=\frac{b}{4}>661. If az>4, then from (1.7), we get ayz-661=\frac{3y+z}{d}<4y. It implies that y\leq (az-4)y<661, which contradicts with y>661. Hence az\leq 4. On the other hand, we have (adz-3)b=az^2+1983\leq 16+1983=1999 which contradicts with b>2644. Similarly we can prove that y\leq z is impossible.
Case 2: 90\leq b\leq 2644, d<7 and a<5, then we have a\in\{1, 2, 3, 4\}, d\in\{1, 2, 4, 5\} since d|3y+z, \gcd(3y, z)=1.
If a=1, from equation (1.7), we get
(dy-1)(dz-3)=661{{d}^{2}}+3. | (4.1) |
Replacing d by 1, 2, 4, 5 in (4.1) respectively, we obtain
(y-1)(z-3)=8\cdot 83, (2y-1)(2z-3)=2647, (4y-1)(4z-3)=71\cdot 149, (5y-1)(5z-3)=16\cdot 1033. |
Thus we get
(d, y, z)=(1, 2, 667), (1, 3, 335), (1, 5, 169), (1, 9, 86), (1, 84, 11), (1, 167, 7), (1, 333, 5), |
(1, 665, 4), (2, 1, 1325), (2, 1324, 2), (4, 2645, 1), (5, 1, 827), (5, 1653, 1). |
If a=2, from equation (1.7), we get
(2dy-1)(2dz-3)=1322{{d}^{2}}+3. | (4.2) |
Replacing d by 1, 2, 4, 5 in (4.2) respectively, we obtain
(2y-1)(2z-3)=25\cdot 53, (4y-1)(4z-3)=11\cdot 13\cdot 17, (8y-1)(8z-3)=5\cdot 4231, (5y-1)(5z-3)=33053. |
Thus we get
\begin{align} & (d, y, z)=(1, 1, 664), (1, 3, 134), (1, 27, 14), (1, 133, 4), (1, 663, 2), \\ & (2, 1323, 1), (4, 529, 1). \\ \end{align} |
If a=3, from equation (1.7), we get
(3dy-1)(dz-1)=661{{d}^{2}}+1. | (4.3) |
Replacing d by 1, 2, 4, 5 in (4.3) respectively, we obtain
(3y-1)(z-1)=2\cdot 331, (6y-1)(2z-1)=5\cdot {{23}^{2}}, (9y-1)(3z-1)=2\cdot {{5}^{2}}\cdot 7\cdot 17, (15y-1)(5z-1)=2\cdot 8263. |
By computing we get that
(d, y, z)=(1, 1, 332), (1, 221, 2), (2, 1, 265), (2, 441, 1). |
If a=4, we contend d<3. Otherwise d\geq 4, we have 3y+z=db\geq 4b. If b\geq 300, then we have yz\geq 1197, which contradicts with yz\leq [\frac{3305}{4}]=826. If 90\leq b<300, then we have yz\geq 357, which contradicts with yz\leq [\frac{961}{4}]=240. Hence d<3 as desired. From equation (1.7), we get
(4dy-1)(4dz-3)=2644{{d}^{2}}+3. | (4.4) |
Replacing d by 1, 2 in (4.4) respectively, we obtain
(4y-1)(4z-3)=2647, (8y-1)(8z-3)=71\cdot 149. |
Thus we get
(d, y, z)=(1, 662, 1). |
Case 3: 90\leq b\leq 2644, d\geq 7 or a\geq 5. Then we have either y\leq 5 or z\leq 5 by Lemma 2.7.
Subcase 1: d=1, a\geq 5, from equation (1.7), we get
\left( ay-1 \right)z=661+3y | (4.5) |
and
\left( az-3 \right)y=661+z. | (4.6) |
Replacing y by 1, 2, 3, 4, 5 in (4.5) respectively, we obtain
(a-1)z=8\cdot 83, (2a-1)z=23\cdot 29, (3a-1)z=2\cdot 5\cdot 67, (4a-1)z=673, (5a-1)z=4\cdot {{13}^{2}}. |
Thus we get
(a, y, z)=(5, 1, 166). |
Replacing z by 1, 2, 4, 5 in (4.6) respectively, we obtain
(a-3)y=2\cdot 331, (2a-3)y=3\cdot 13\cdot 17, (4a-3)y=5\cdot 7\cdot 19, (5a-3)y=2\cdot {{3}^{2}}\cdot 37. |
By computing we get
(a, y, z)=(5, 331, 1), (8, 51, 2), (10, 39, 2). |
Subcase 2: d=2, a\geq 5, from equation (1.7), we get
\left( 2ay-1 \right)z=1322+y | (4.7) |
and
\left( 2az-3 \right)y=1322+z. | (4.8) |
Replacing y by 1, 2, 3, 4, 5 in (4.7) respectively, we obtain
(2a-1)z={{5}^{2}}\cdot 53, (4a-1)z={{2}^{4}}\cdot 83, (6a-1)z={{11}^{3}}, (8a-1)z=2\cdot 23\cdot 29, (10a-1)z=7\cdot 191. |
By computing we know that the above equations have no positive integer solution (a, z) such that a\geq 5, b\geq 90.
Replacing z by 1, 2, 4, 5 in (4.8) respectively, we obtain
(2a-3)y={{3}^{3}}\cdot {{7}^{2}}, (4a-3)y={{2}^{2}}\cdot 331, (8a-3)y=2\cdot 3\cdot 13\cdot 17, (10a-3)y=1327. |
By computing we get
(a, y, z)=(6, 147, 1), (12, 63, 1), (5, 189, 1). |
Subcase 3: d\geq 4, a\geq 5, then we have 3y+z=db\geq 4b. If b\geq 200, then we get that yz\geq 797 which contradicts with yz\leq [\frac{3305}{5}]=661. If 90\leq b<200, then we get that yz\geq 357 which contradicts with yz\leq [\frac{861}{5}]=172.
Subcase 4: d\geq 7, a<5. We contend that a=1. Otherwise a\geq 2. If b\geq 300, then we get that yz\geq 2097 which contradicts with yz\leq [\frac{3305}{2}]=1652. If 90\leq b<300, then we get that yz\geq 627 which contradicts with yz\leq [\frac{861}{2}]=430. Thus a=1 as desired. From equation (1.7), we get
(dy-1)(yz-661)=661+3{{y}^{2}} | (4.9) |
and
(dz-3)(yz-661)=1983+{{z}^{2}}. | (4.10) |
Replacing y by 1, 2, 3, 4, 5 in (4.9) respectively, we obtain
\begin{align} & (d-1)(z-661)=8\cdot 83, (2d-1)(2z-661)=673, (3d-1)(3z-661)={{4}^{2}}\cdot 43, (4d-1)(4z-661)=709, \\ & (5d-1)(5z-661)={{2}^{5}}\cdot 23. \\ \end{align} |
By computing we know that the above equations have no positive integer solution (a, z) such that d\geq 7, b\geq 90.
Replacing z by 1, 2, 4, 5 in (4.8) respectively, we obtain
(d-3)(y-661)={{2}^{6}}\cdot 31, (2d-3)(2y-661)=1987, (4d-3)(4y-661)=1999, (5d-3)(5y-661)=8\cdot 251. |
By computing we get that
(d, y, z)=(7, 1157, 1), (11, 909, 1), (19, 785, 1). |
Therefore the proof of Theorem 1.2 is complete.
Since
\begin{align} & \begin{matrix} 661+12=673\equiv 1 & \left( \bmod 3 \right), 661+22=683\equiv 1 & \left( \bmod 11 \right), 661+30=691\equiv 1 & \left( \bmod 3 \right), \\ \end{matrix} \\ & \begin{matrix} 661+40=701\equiv 1 & \left( \bmod 5 \right), 661+48=709\equiv 1 & \left( \bmod 3 \right), 661+58=719\equiv 23 & \left( \bmod 29 \right), \\ \end{matrix} \\ & \begin{matrix} 661+66=727\equiv 1 & \left( \bmod 3 \right), 661+72=733\equiv 1 & \left( \bmod 3 \right), 661+78=739\equiv 1 & \left( \bmod 3 \right), \\ \end{matrix} \\ & \begin{matrix} 661+82=743\equiv 5 & \left( \bmod 41 \right), \\ \end{matrix} \\ \end{align} |
so both (1.6) and (1.7) have no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{12, 22, 30, 40, 48, 58, 66, 72, 78, 82\} by Lemma 2.1.
Since
\begin{align} & \begin{matrix} 661+10=671=11\cdot 61, 11\equiv 61\equiv 1 & \left( \bmod 5 \right), \\ \end{matrix} \\ & \begin{matrix} 661+20=681=3\cdot 227, 227\equiv 7 & \left( \bmod 20 \right), 661+26=687=3\cdot 229, 229\equiv -5 & \left( \bmod 13 \right), \\ \end{matrix} \\ & 661+28=689=13\cdot 53, \\ & \begin{matrix} 661+34=695=5\cdot 139, 139\equiv 3 & \left( \bmod 17 \right), \\ \end{matrix} \\ & \begin{matrix} 661+37=698=2\cdot 347, 347\equiv 14 & \left( \bmod 37 \right), \\ \end{matrix} \\ & \begin{matrix} 661+38=699=3\cdot 233, 233\equiv 5 & \left( \bmod 19 \right), \\ \end{matrix} \\ & \begin{matrix} 661+46=707=7\cdot 101, 101\equiv 9 & \left( \bmod 23 \right), \\ \end{matrix} \\ & 661+52=713=23\cdot 31, \\ & \begin{matrix} 661+56=717=3\cdot 239, 239\equiv 1 & \left( \bmod 7 \right), \\ \end{matrix} \\ & \begin{matrix} 661+57=718=2\cdot 359, 359\equiv 17 & \left( \bmod 57 \right), \\ \end{matrix} \\ & \begin{matrix} 661+60=721=7\cdot 103, 7\equiv 103\equiv 1 & \left( \bmod 3 \right), \\ \end{matrix} \\ & \begin{matrix} 661+62=723=3\cdot 241, 241\equiv -7 & \left( \bmod 31 \right), \\ \end{matrix} \\ & \begin{matrix} 661+70=731=17\cdot 43, 17\equiv 3 & \left( \bmod 7 \right), 43\equiv 1 & \left( \bmod 7 \right), \\ \end{matrix} \\ & \begin{matrix} 661+56=717=3\cdot 239, 239\equiv 15 & \left( \bmod 56 \right), \\ \end{matrix} \\ & \begin{matrix} 661+73=734=2\cdot 367, 367\equiv 2 & \left( \bmod 73 \right), \\ \end{matrix} \\ & 661+76=737=11\cdot 67, \\ & \begin{matrix} 661+85=746=2\cdot 373, 373\equiv 33 & \left( \bmod 85 \right), \\ \end{matrix} \\ & \begin{matrix} 661+88=749=7\cdot 107, 107\equiv 19 & \left( \bmod 88 \right), \\ \end{matrix} \\ \end{align} |
so both (1.6) and (1.7) have no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{10, 20, 26, 28, 34, 37, 38, 46, 52, 56, 57, 60, 62, 70, 73, 76, 85, 88\} by Lemma 2.2 (i).
Since
\begin{align} & \begin{matrix} 661+31=692={{2}^{2}}\cdot 173, 173\equiv 18 & \left( \bmod 31 \right), \\ \end{matrix} \\ & \begin{matrix} 661+44=705=3\cdot 5\cdot 47, 47\equiv 3 & \left( \bmod 11 \right), \\ \end{matrix} \\ & \begin{matrix} 661+50=711={{3}^{2}}\cdot 79, 79\equiv 4 & \left( \bmod 25 \right), \\ \end{matrix} \\ & \begin{matrix} 661+55=716={{2}^{2}}\cdot 179, 179\equiv 14 & \left( \bmod 55 \right), \\ \end{matrix} \\ & 661+61=722=2\cdot {{19}^{2}}, \\ & 661+64=725={{5}^{2}}\cdot 29, \\ & 661+80=741=3\cdot 13\cdot 19, \\ \end{align} |
so both (1.6) and (1.7) have no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{31, 44, 50, 55, 61, 64\} by Lemma 2.3 (i).
Since
\begin{align} & 661+47=708={{2}^{2}}\cdot 3\cdot 59, 661+59=720={{2}^{4}}\cdot {{3}^{2}}\cdot 5, 661+68=729={{3}^{6}}, \\ & 661+71=732={{2}^{2}}\cdot 3\cdot 61, 661+77=738=2\cdot {{3}^{2}}\cdot 41, \\ \end{align} |
by computing we know both (1.6) and (1.7) have no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{47, 59, 68, 71, 77\}.
Since 661+27=688=2^4\cdot43, 27\nmid x, for any
x\in \{1+{{2}^{\alpha }}{{43}^{\beta }}, 43+{{2}^{\alpha }}, {{2}^{t}}+{{43}^{\beta }}\}, |
where \alpha\in\{0, 1, 2, 3, 4\}, \beta\in\{0, 1, \}, t\in\{1, 2, 3, 4\}.661+75=736=2^5\cdot23, 75\nmid x, for any
x\in \{1+{{2}^{\alpha }}{{23}^{\beta }}, 23+{{2}^{\alpha }}, {{2}^{t}}+{{23}^{\beta }}\}, |
where \alpha\in\{0, 1, 2, 3, 4, 5\}, \beta\in\{0, 1, \}, t\in\{1, 2, 3, 4, 5\} so (1.6) has no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{27, 75\} by Lemma 2.5.
Since
\begin{align} & \begin{matrix} 661+18=679=7\cdot 97, 7\equiv 97\equiv -2 & \left( \bmod 9 \right), \\ \end{matrix} \\ & \begin{matrix} 661+24=685=5\cdot 137, 5\equiv 137\equiv 1 & \left( \bmod 4 \right), \\ \end{matrix} \\ & \begin{matrix} 661+33=694=2\cdot 347, 347\equiv 6 & \left( \bmod 11 \right), \\ \end{matrix} \\ & \begin{matrix} 661+36=697=17\cdot 41, 17\equiv 41\equiv 1 & \left( \bmod 4 \right), \\ \end{matrix} \\ & \begin{matrix} 661+42=703=19\cdot 37, 19\equiv 37\equiv 1 & \left( \bmod 3 \right), \\ \end{matrix} \\ & \begin{matrix} 661+45=706=2\cdot 353, 353\equiv 2 & \left( \bmod 9 \right), \\ \end{matrix} \\ \end{align} |
so (1.6) has no positive integer solution (a, d, y, z) with ayz-661=b for b\in\{18, 24, 33, 36, 42, 45\} by Lemma 2.2 (ii).
For any b=3t, 1\leq t<29, (1.7) has no positive integer solution (a, d, y, z) with ayz-661=b by Lemma 2.4.
Since
\begin{align} & 661+3={{2}^{3}}\cdot 83, 661+6=2\cdot {{3}^{2}}\cdot 37, 661+9=2\cdot 5\cdot 67, 661+15={{2}^{2}}\cdot {{13}^{2}}, \\ & \\ \end{align} |
661+21=2\cdot 11\cdot 31, 661+39={{2}^{2}}\cdot {{5}^{2}}\cdot 7, 661+51={{2}^{3}}\cdot 89, |
so by computing we get that (a, d, y, z)=
(2, 29, 4, 83), (2, 112, 1, 332), (8, 28, 1, 83), (83, 3, 1, 8), (332, 1, 1, 2), (23, 5, 1, 29),
(29, 4, 1, 23), (2, 8, 5, 67), (5, 15, 1, 134), (26, 1, 2, 13), (2, 2, 11, 31), (11, 3, 1, 62),
(2, 1, 14, 25), (5, 1, 4, 35), (2, 7, 1, 356),
are positive integer solutions of (1.6).
We are now in a position to discuss
b\in \{1, 2, 4, 5, 7, 8, 11, 13, 14, 16, 17, 19, 23, 25, 28, 29, 32, 35, 41, 43, 49, 53, 59, 65, 67, 74, 79\}. |
Since
\begin{align} & 661+1=2\cdot 331, 661+2=3\cdot 13\cdot 17, 661+4=5\cdot 7\cdot 19, 661+5=2\cdot {{3}^{2}}\cdot 37, \\ & 661+7={{2}^{2}}\cdot 167, 661+8=3\cdot 223, 661+11={{2}^{5}}\cdot 3\cdot 7, 661+13=2\cdot 337, 661+14={{3}^{3}}\cdot {{5}^{2}} \\ & 661+16=677, 661+17=2\cdot 3\cdot 113, 661+19={{2}^{3}}\cdot 5\cdot 17, 661+23={{2}^{2}}\cdot {{3}^{2}}\cdot 19, \\ & 661+25=2\cdot {{7}^{3}}, 661+28=13\cdot 53, 661+29=2\cdot 3\cdot 5\cdot 23, 661+32={{3}^{2}}\cdot 7\cdot 11, \\ & 661+35={{2}^{3}}\cdot 3\cdot 29, 661+41=2\cdot {{3}^{3}}\cdot 13, 661+43={{2}^{6}}\cdot 11, 661+49=2\cdot 5\cdot 71, \\ & 661+53=2\cdot 3\cdot 7\cdot 17, 661+59={{2}^{4}}\cdot 5\cdot {{3}^{2}}, 661+65=2\cdot 3\cdot {{11}^{2}}, \\ & 661+67={{2}^{3}}\cdot 7\cdot 13, 661+74=3\cdot 5\cdot {{7}^{2}}, 661+79={{2}^{2}}\cdot 5\cdot 37, \\ \end{align} |
so by computing we get that (a, d, y, z)=
(662, 2, 1, 1), (1, 663, 1, 662), (1, 333, 2, 331), (2, 332, 1, 331), (331, 3, 1, 2), (663, 1, 1, 1),
(1, 332, 1, 663), (1, 112, 3, 221), (1, 32, 13, 51), (1, 28, 17, 39), (3, 111, 1, 221), (3, 14, 13, 17),
(13, 26, 1, 51), (13, 10, 3, 17), (17, 20, 1, 39), (17, 8, 3, 13), (39, 9, 1, 17), (51, 7, 1, 13),
(221, 2, 1, 3), (7, 6, 5, 19), (19, 3, 5, 7), (7, 24, 1, 95), (19, 9, 1, 35), (35, 5, 1, 19),
(95, 2, 1, 7), (1, 11, 18, 37), (1, 67, 2, 333), (9, 15, 1, 74), (74, 2, 1, 9), (6, 8, 3, 37),
(111, 1, 2, 3), (4, 24, 1, 167), (3, 28, 1, 223), (32, 2, 1, 21), (24, 1, 4, 7), (28, 1, 3, 8),
(21, 3, 1, 32), (2, 26, 1, 337), (15, 1, 5, 9), (9, 2, 3, 25), (25, 2, 1, 27), (1, 7, 6, 113),
(2, 20, 1, 339), (1, 3, 17, 40), (20, 1, 2, 17), (9, 1, 4, 19), (14, 2, 1, 49), (5, 1, 6, 23),
(6, 4, 1, 115), (11, 2, 1, 63), (3, 1, 11, 21), (4, 1, 6, 29), (4, 5, 1, 174), (9, 1, 2, 39),
(2, 1, 11, 32), (1, 3, 5, 142), (7, 1, 2, 51), (5, 2, 1, 147), (2, 1, 5, 74)
are the solutions of (1.6) and (a, d, y, z)=
(662, 4, 1, 1), (1, 337, 2, 331), (1, 665, 1, 662), (1, 1987, 662, 1), (1, 995, 331, 2), (2, 334, 1, 331),
(2, 994, 331, 1), (331, 5, 1, 2), (331, 7, 2, 1), (663, 2, 1, 1), (1, 333, 1, 663), (1, 995, 663, 1),
(1, 115, 3, 221), (1, 83, 51, 13), (1, 67, 39, 17), (3, 112, 1, 221), (3, 332, 221, 1), (3, 28, 13, 17),
(3, 32, 17, 13), (13, 13, 3, 17), (17, 59, 39, 1), (17, 11, 3, 13), (39, 10, 1, 17), (39, 26, 17, 1),
(51, 8, 1, 13), (51, 20, 13, 1), (221, 5, 3, 1), (1, 167, 1, 665), (665, 1, 1, 1), (1, 37, 5, 133),
(1, 101, 133, 5), (1, 31, 35, 19), (1, 73, 95, 7), (5, 34, 1, 133), (5, 10, 7, 19), (5, 16, 19, 7),
(133, 2, 1, 5), (5, 100, 133, 1), (333, 1, 1, 2), (18, 8, 1, 37), (222, 2, 3, 1), (3, 11, 6, 37),
(3, 67, 111, 2), (37, 11, 18, 5), (2, 200, 333, 1), (334, 1, 2, 1), (167, 1, 1, 4), (1, 29, 3, 223),
(6, 5, 16, 7), (8, 23, 84, 1), (84, 1, 1, 8), (96, 2, 7, 1), (112, 1, 3, 2), (45, 1, 3, 5),
(27, 2, 1, 25), (75, 2, 9, 1), (1, 127, 677, 1), (3, 7, 2, 113), (6, 20, 113, 1), (34, 1, 5, 4),
(4, 2, 9, 19), (13, 2, 1, 53), (15, 1, 2, 23), (9, 1, 7, 11), (33, 2, 21, 1), (12, 1, 2, 29),
(12, 5, 58, 1), (3, 2, 4, 58), (27, 1, 13, 2), (26, 2, 27, 1), (8, 1, 18, 5), (2, 2, 3, 121),
(2, 1, 13, 28), (15, 2, 49, 1)
are the solutions of (1.7).
We are now in a position to discuss b\in\{83, 86, 89\}.
Since 661+83=2^3\cdot3\cdot31, 661+86=3^2\cdot83 so by computing we get that (a, d, y, z)=(1, 9, 1, 744), (9, 1, 1, 83) are solutions of (1.7).
Since 661+89=2\cdot3\cdot5^3, so by computing we get that (1.7) has no positive integer solution.
This finishes the proof of Theorem 1.3.
If ay>2, then from ayz-661=\frac{661y+z}{d}<2z, we get that z\leq (ay-2)z<661 which contradicts with z>661y. Hence a=y=1 or a=1, y=2 or a=2, y=1.
a=y=1 implies that (d-1)(z-661)=2\cdot661. So
(d, z)=(2, 1983), (3, 1322), (662, 663), (1323, 662). |
a=1, y=2 implies that (2d-1)(2z-661)=5\cdot661. So
(d, z)=(1, 1983), (3, 661), (331, 333), (1653, 331). |
a=2, y=1 implies that (2d-1)(2z-661)=3\cdot661. So
(d, z)=(1, 1322), (2, 661), (331, 332), (992, 331). |
This concludes the proof of Theorem 1.4.
All authors declare no conflicts of interest in this paper.
[1] | Z. Cao, An intruduction to Diophantine equations, Haerbin Industril university Press, 400 (1989). |
[2] | N. Franceschine, Egyptian fractions, Sonoma State Coll., 1978. |
[3] | Z. Ke, Q. Sun, X. Zhang, On the Diophantine equatins \frac{4}{n}=\frac{1}{x}+\frac{1}{y}+\frac{1}{z}, J. Sichuan University (in Chinese), 3 (1964), 23-37. |
[4] | Z. Ke, Q. Sun, Some conjecture and problem in Number Theorey, Chinese Journal of Nature (in Chinese), 7 (1979), 411-413. |
[5] | Y. Liu, On a problem of unit fraction, J. Sichuan University (in Chinese), 2 (1984), 113-114. |
[6] | Palama, Sull’s equarione diofantea \frac{4}{n}=\frac{1}{x}+\frac{1}{y}+\frac{1}{z}, Boll. Union Mat. Ital, 13 (1958), 65-72. |
[7] | Palama, Sull’s equarione diofantea \frac{4}{n}=\frac{1}{x}+\frac{1}{y}+\frac{1}{z}, Boll. Union Mat. Ital, 14 (1959), 82-94. |
[8] | L. A. Rosati, Sull’s equarione diofantea \frac{4}{b}=\frac{1}{x_1}+\frac{1}{x_2}+\frac{1}{x_3}, Boll. Union Mat. Ital, 9 (1954), 59-63. |
[9] | B.M.Stewart, Theory of numbers, Macmillan, New York, (1964), 198-207. |
[10] | Yamamoto, On the Diophantine equatins \frac{4}{n}=\frac{1}{x}+\frac{1}{y}+\frac{1}{z}, K. Men. Fac. Sci. Kyushu University, Ser.A., 19 (1965), 37-47. |