This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Steps in Anderson-Badawi’s Conjecture on n-Absorbing and Strongly n-Absorbing Ideals

Matija Delic Department of Mathematics, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA [email protected]  and  Khalid Adarbeh (⋆) Department of Mathematics, An-Najah National University, Nablus, Palestine [email protected]
Abstract.

This article aims to solve positively Anderson-Badawi Conjecture of n-Absorbing and strongly n-absorbing ideals of commutative rings in the class of u-rings. The main result extends and recovers Anderson-Badawi’s related result on Prufer domains [2, Corollary 6.9].

Key words and phrases:
2-absorbing ideals, n-absorbing ideals, Strongly 2-absorbing ideals, u-rings.
2010 Mathematics Subject Classification:
13A15, 13F05, 13G05.
(⋆) Corresponding author

1. Introduction

Throughout this article, RR denotes a commutative ring with 101\neq 0. In 2007, A. Badawi introduced the concept of 2-absorbing ideals of commutative rings as a generalization of prime ideals. He defined an ideal II of RR to be 2-absorbing if whenever a,b,cRa,b,c\in R and abcIabc\in I, then abab or acac or bcbc is in II [3]. As in the case of prime ideals, 2-absorbing have a characterization in terms of ideals. Namely, II is 2-absorbing if whenever I1,I2,I3I_{1},I_{2},I_{3} are ideals of RR and I1I2I3II_{1}I_{2}I_{3}\subseteq I, then I1I2I_{1}I_{2} or I1I3I_{1}I_{3} or I2I3I_{2}I_{3} is contained in II [3, Theorem 2.13].

In 2011, D.F. Anderson, A. Badawi inspired from the definition of 2-absorbing ideals and defined the n-absobing ideals for any positive integer n. Where an ideal II is called n-absorbing ideal if whenever x1xn+1Ix_{1}\dots x_{n+1}\in I for x1,,xn+1Rx_{1},\dots,x_{n+1}\in R then there are n of the xix_{i}’s whose product is in I. Also they introduced the strongly-n-absorbing ideals as another generalization of prime ideals, where an ideal II of RR is said to be a strongly n-absorbing ideal if whenever I1···In+1II_{1}\dots\textperiodcentered\textperiodcentered\textperiodcentered I_{n+1}\subseteq I for ideals I1,,···,In+1I_{1},\dots,\textperiodcentered\textperiodcentered\textperiodcentered,I_{n+1} of RR, then the product of some n of the IjI_{j}’s is contained in II. Obviously, a strongly n- absorbing ideal of RR is also an n-absorbing ideal of RR, and by the last fact in the previous paragraph, 2-absorbing and strongly 2 absorbing are the same. Moreover D.F. Anderson, A. Badawi were able to prove that n-absorbing and strongly n-absorbing are equivalent in the class of Prufer domains [2, Corollary 6.9], and they conjectured that these two concepts are equivalent in any commutative ring [2, Conjecture 1].

In 1975, Jr. P. Quartararo and H.S. Butts defined the u-rings to be those rings in which if an ideal II is contained in the union of ideals, then it must be contained in one of them. Then they proved that it suffices to consider the case II is finitely generated ideal of RR [5, Proposition 1.1]. Moreover, in [5, Corollary 1.6], they proved that the class of Prufer domains (domains in which every finitely generated ideal is invertible) is contained in the class of u-rings. So we have the following diagram of implications:

Prufer domains

\Downarrow

u-rings

where the implication is irreversible in general; see Example 3.9 for a u-ring which is not a domain, particularly, not a Prufer domain.

In section one of this paper, we provide an alternative proof of [3, Theorem 2.13]. The technique of this proof helps in proving the main result of Section 2. In section 2, we solve positively Anderson-Badawi’s Conjecture of n-Absorbing and strongly n-absorbing ideals in the class of u-rings. The main result (Theorem 3.1) extends and recovers Anderson-Badawi’s related result on Prufer domains (Corollary 3.7).

2. Alternative proof

As we mentioned in the introduction, 2-absorbing ideals and strongly 2-absorbing are the same. This follows trivially from [3, Theorem 2.13]. In this section, we present an alternative proof of [3, Theorem 2.13], which inspires us in solving [2, Conjecture 1] in the class of u-rings. For the seek of completeness, We provide the proof of the following lemma; which can be found as an exercise in the classical ring theory texts.

Lemma 2.1.

Let II be an ideal of RR. If I=I1I2I=I_{1}\cup I_{2}, where I1I_{1} and I2I_{2} are also ideals, then I=I1I=I_{1} or I=I2I=I_{2}.

Proof.

Suppose I1I2I_{1}\setminus I_{2} and I2I1I_{2}\setminus I_{1} are nonempty. Let aI1I2a\in I_{1}\setminus I_{2} and bI2I1b\in I_{2}\setminus I_{1}. Since I1I2I_{1}\cup I_{2} is ideal, a+bI1I2a+b\in I_{1}\cup I_{2}. Assume, without loss of generality, that a+bI1a+b\in I_{1}. Then b=(a+b)aI1b=(a+b)-a\in I_{1}, a contradiction. Therefore, either I1I2=ϕI_{1}\setminus I_{2}=\phi or I2I1=ϕI_{2}\setminus I_{1}=\phi; equivalently, I1I2I_{1}\subseteq I_{2} or I2I1I_{2}\subseteq I_{1}. So that I=I1I=I_{1} or I=I2I=I_{2}. ∎

Now, we prove a few lemmas in a sequence, finishing with the proof of the theorem.

Lemma 2.2.

Suppose that II is a 2-absorbing ideal of RR, JJ is an ideal of RR and xyJIxyJ\subseteq I for some x,yRx,y\in R. Then xyIxy\in I or xJIxJ\subseteq I or yJIyJ\subseteq I.

Proof.

Suppose xyIxy\not\in I. Denote by Jx={zJ|xzI}J_{x}=\{z\in J\ |\ xz\in I\} and Jy={zJ|yzI}J_{y}=\{z\in J\ |\ yz\in I\}. It is not hard to show that JxJ_{x} and JyJ_{y} are ideals. Now, if aJa\in J, then xyaIxya\in I. But II being 2-absorbing and xyIxy\not\in I imply that xaIxa\in I or yaIya\in I. Thus, either aJxa\in J_{x} or aJya\in J_{y}, and hence J=JxJyJ=J_{x}\cup J_{y}. Therefore, by Lemma 2.1, either J=JxJ=J_{x}, and hence xJIxJ\subseteq I or J=JyJ=J_{y}, and hence yJIyJ\subseteq I. ∎

We generalize the previous lemma as follows.

Lemma 2.3.

Suppose that II is a 2-absorbing ideal of RR, I1I_{1} and I2I_{2} are ideals of RR, and xI1I2IxI_{1}I_{2}\subseteq I for some xRx\in R. Then xI1IxI_{1}\subseteq I or xI2IxI_{2}\subseteq I or I1I2II_{1}I_{2}\subseteq I.

Proof.

Suppose xI2IxI_{2}\not\subseteq I. By Lemma 2.2, for all yI1y\in I_{1}, either xyIxy\in I or yI2IyI_{2}\subseteq I. Let N={yI1|xyI}N=\{y\in I_{1}\ |\ xy\in I\} and M={yI1|yI2I}M=\{y\in I_{1}\ |\ yI_{2}\subseteq I\}. Then MM and NN are ideals of RR, and simlirly as in the proof of Lemma 2.2, I1=NMI_{1}=N\cup M. Thus, again by Lemma 2.1, either I1=NI_{1}=N, and in this case xI1IxI_{1}\subseteq I, or I1=MI_{1}=M, and in this case I1I2II_{1}I_{2}\subseteq I. ∎

Finally, we use the above lemmas to prove the main theorem.

Theorem 2.4.

[3, Theorem 2.13] An ideal II of RR is 2-absorbing ideal if and only if it is strongly 2-absorbing ideal.

Proof.

Obviously, strongly 2-absorbing ideals are 2-absorbing. Conversely, Assume that II is 2-absorbing and I1I2I3II_{1}I_{2}I_{3}\subseteq I, where I1I_{1}, I2I_{2}, and I3I_{3} are ideals of RR. Further, Suppose I2I3II_{2}I_{3}\not\subseteq I, and let N={xI1|xI2I}N=\{x\in I_{1}\ |\ xI_{2}\subseteq I\} and M={xI1|xI3I}M=\{x\in I_{1}\ |\ xI_{3}\subseteq I\}. Then MM and NN are ideals. By Lemma 2.3, all xI1x\in I_{1} are in either NN or MM, and thus I1=NMI_{1}=N\cup M. Therefore by Lemma 2.1, either I=NI=N or I=MI=M; which implies that I1I2II_{1}I_{2}\subseteq I or I1I3II_{1}I_{3}\subseteq I. ∎

3. The conjecture

The following conjecture was announced in [2].
Anderson and Badawi’s conjecture: In every ring, the notions of nn-absorbing ideals and strongly nn-absorbing ideals are equivalent.

It is easy to see that strongly nn-absorbing ideals are nn-absorbing. We aim to find conditions for the converse to hold. We adopt the following terminology from [8] and [5]: If I1,,InI_{1},...,I_{n} are ideals of RR, then I1InI_{1}\cup...\cup I_{n} is called an efficient covering of II if II1InI\subseteq I_{1}\cup...\cup I_{n}, but II is not contained in the union of any n1n-1 of these ideals [8]. In view of this definition, an ideal II of RR is called a u-ideal if there is no efficient covering of II with more then one ideal.

The following result solves Anderson and Badawi’s conjecture to u-rings, generalizing thus Corollary 6.9 from [2].

Theorem 3.1.

In a uu-ring, an nn-absorbing ideal is strongly nn-absorbing.

In order to prove this main theorem, we prove the following four lemmas:

Lemma 3.2.

A principal ideal is a u-ideal.

Proof.

Say II1InI\subseteq I_{1}\cup...\cup I_{n}, and I=(x)I=(x). Then for some jj, xIjx\in I_{j} so IIjI\subseteq I_{j}. ∎

Lemma 3.3.

Let II be an nn-absorbing ideal of RR, and I1,,In+1I_{1},...,I_{n+1} be uu-ideals of RR. Suppose that the following condition is satisfied:

whenever I1In+1II_{1}\cdots I_{n+1}\subseteq I, and at least k+1k+1 of the ideals I1,,In+1I_{1},...,I_{n+1} are principal, then II contains a product of some nn of them.

Then the same holds when we replace k+1k+1 with kk. Here nk0n\geq k\geq 0.

Proof.

Assume the statement is true for II and k+1k+1. Let I1In+1II_{1}\cdots I_{n+1}\subseteq I, where IjI_{j} is principal for jkj\leq k. Assume jnIjI\prod_{j\leq n}I_{j}\not\subseteq I. For all ini\leq n, let

Ji={yIn+1|yjn+1,iIjI}J_{i}=\{y\in I_{n+1}\ |\quad y\prod_{j\neq n+1,i}I_{j}\subseteq I\}

Then by our assumption, In+1=inJiI_{n+1}=\cup_{i\leq n}J_{i}. Since it is a uu-ideal, it is equal to some JiJ_{i}. But then

jiIjI\prod_{j\neq i}I_{j}\subseteq I

This concludes the proof. ∎

Lemma 3.4.

Let II be an nn-absorbing ideal. If I1In+1II_{1}\cdots I_{n+1}\subseteq I, where every IjI_{j} is a uu-ideal, then II contains the product of some nn of these ideals.

Proof.

By the definition of II, and Lemma 3.2, the statement holds when I1,,In+1I_{1},...,I_{n+1} are all principal ideals. We use Lemma 3.3 to induct down from the case k=nk=n (where we require k+1k+1 ideals to be principle) to k=0k=0 (where we require no ideals to be principle), which is exactly what we want. ∎

This allows us to prove the main theorem of this article (Theorem 3.1).

Proof of Theorem 3.1: Assume the contrary. Then in some uu-ring, there are ideals I,I1,,In+1I,I_{1},...,I_{n+1} such that II is nn-absorbing and I1In+1II_{1}\cdots I_{n+1}\subseteq I, but II doesn’t contain the product of any nn of these ideals. But RR is a u-ring, and hence I1,,InI_{1},...,I_{n} are uu-ideals. Lemma 3.4 gives a contradiction.

Remark 3.5.

We can alter the proof of Lemma 3.4 above slightly, to get a more general statement when n=2n=2. Indeed, notice that if I=I1I2I=I_{1}\cup I_{2}, then I=I1I=I_{1} or I2I_{2} (well-known). Then we can drop the condition of the ideals needing to be uu-ideals from Lemma 3.4, and hence we obtain for arbitrary rings, every 22-absorbing ideal is strongly 22-absorbing. This is Theorem 2.4.

We can use this to give an alternative proof to corollary 6.9 from [2]. To achieve that, we cite the following results first.

Proposition 3.6.

Every invertible ideal is a uu-ideal, and a Prüfer domain is a uu-ring.

Proof.

See Theorem 1.5 and Corollary 1.6 from [5]. ∎

As a straightforward application of Theorem 3.1, we recover Anderson-Badawi’s related result on Prufer domains

Corollary 3.7.

In Prüfer domains, an nn-absorbing ideal is strongly nn-absorbing.

Lastly, to ensure that u-rings is strictly larger that the class of Prüfer domains, we prove the following lemma which provides an example of one such family of u-rings. A more general result, proved in the same way, can be found in [5].

Lemma 3.8.

Suppose RR is a ring with R\mathbb{Q}\subseteq R. Then RR is a uu-ring.

Proof.

Let I=I1InI=I_{1}\cup\dots\cup I_{n} be an efficient covering of II. Take a1I1a_{1}\in I_{1} with a1Ija_{1}\not\in I_{j} for j1j\neq 1. Choose a2a_{2} analogously. Then for all kk\in\mathbb{Z}, a1+ka2I1,I2a_{1}+ka_{2}\not\in I_{1},I_{2}. Since there are infinite possibilities for kk, there will be a1+ka2a_{1}+ka_{2} and a1+la2a_{1}+la_{2} in the same IjI_{j}. But then (kl)a2Ij(k-l)a_{2}\in I_{j}, so a2Ija_{2}\in I_{j} for j2j\neq 2, contradiction. ∎

The following is an example of a u-ring which is not a domain, and hence not a Prüfer domain.

Example 3.9.

×\mathbb{Q}\times\mathbb{Q} is a ring with zero divisors (not domain) which contains 0×\mathbb{Q}\cong{0}\times\mathbb{Q} as a subring. Consequently, by Lemma 3.8, ×\mathbb{Q}\times\mathbb{Q} is a u-ring.

Acknowledgements. We are grateful to the Undergraduate Research Opportunities Program at MIT (UROP) as well as to the J-WEL Grant in Higher Education Innovation, “Educational Innovation in Palestine,” for providing and funding this research opportunity. Also, we would like to thank Professor Haynes Miller for his crucial role in mentoring this project.

References

  • [1]
  • [2] D.F. Anderson, A. Badawi, On n-absorbing ideals of commutative rings, Comm. Algebra, 39 (2011), 1646-1672.
  • [3] A. Badawi, On 2-absorbing ideals of commutative rings, Bull. Austral. Math. Soc., 75 (2007), 417-429.
  • [4] Yousefian Darani, A., Puczylowski, E.R., On 2-absorbing commutative semigroups and thier applications to rings, Semigroup Forum 86 (2013), 83-91
  • [5] P. Quartararo Jr. & H.S. Butts, Finite unions of ideals and modules, Proc. Amer. Math. Soc. 52 (197W5), 91-96.
  • [6] S. McAdam, Finite coverings by ideals, Ring theory, Proceedings of the Oklahoma conference, Lecture Notes in Pure and Applied Mathematics volume 7, Dekker 1974, 163-171.
  • [7] N. McCoy, A note on finite unions of ideals and subgroups, Proc. Amer.Math. Soc. 8 (1957), 633-637
  • [8] Gottlieb, Christian. On finite unions of ideals and cosets. Communications in Algebra. 22 (1994), 3087-3097