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


Characterizations of II-semiregular and II-semiperfect rings

Yongduo Wang
Department of Applied Mathematics, Lanzhou University of Technology

Lanzhou 730050, P. R. China

E-mail: ydwang@lut.cn


Abstract Let MM be a left module over a ring RR and II an ideal of RR. We call (P,f)(P,f) a (locally)projective II-cover of MM if ff is an epimorphism from PP to MM, PP is (locally)projective, KerfIPKerf\subseteq IP, and whenever P=Kerf+XP=Kerf+X, then there is a projective summand YY of PP in KerfKerf such that P=YXP=Y\oplus X. This definition generalizes (locally)projective covers. We characterize II-semiregular and II-semiperfect rings which are defined by Yousif and Zhou [19] using (locally)projective II-covers in section 2 and 3. II-semiregular and II-semiperfect rings are characterized by projectivity classes in section 4. Finally, the notion of II-supplemented modules are introduced and II-semiregular and II-semiperfect rings are characterized by II-supplemented modules. Some well known results are obtained as corollaries.

Keywords: Semiregular, Semiperfect, Locally projective modules, Projectivity class


1. Introduction and Preliminaries

It is well known that (locally) projective covers, projectivity classes and supplemented modules play important roles in characterizing semiperfect and semiregular rings. Recently, some authors had worked with various extensions of these rings (see for examples [2, 9, 11, 19, 20]). As generalizations of semiregular rings and semiperfect rings, the notions of II-semiregular rings and II-semiperfect rings were introduced by Yousif and Zhou [19]. Our purposes of this paper as follows: (1) characterize II-semiregular and II-semiperfect rings by defining (locally) projective II-covers in section 2 and 3; (2) characterize II-semiregular rings and II-semiperfect rings in term of projectivity classes of modules in section 4. (3) chracterize II-semiregular rings and II-semiperfect rings by defining II-supplemented modules in section 5.

Let RR be a ring and II an ideal of RR, MM a module and SMS\leq M. SS is called small in MM (notation SMS\ll M) if MS+TM\neq S+T for any proper submodule TT of MM. As a proper generalization of small submodules, the concept of δ\delta-small submodules was introduced by Zhou in [20]. Let NMN\leq M. NN is said to be δ\delta-small in MM if, whenever N+X=MN+X=M with M/XM/X singular, we have X=MX=M. δ(M)=RejM()={NMM/N}\delta(M)=Rej_{M}(\wp)=\cap\{N\leq M\mid M/N\in\wp\} , where \wp be the class of all singular simple modules. We also recall that a pair (P,f)(P,f) is called a (locally) projective (δ)(\delta-)cover of MM if PP is (locally) projective, ff is an epimorphism from PP to MM such that KerfPKerf\ll P(KerfδPKerf\ll_{\delta}P). An element mm of MM is called II-semiregular [2] if there exists a decomposition M=PQM=P\oplus Q where PP is projective, PRmP\subseteq Rm and RmQIMRm\cap Q\subseteq IM. MM is called an II-semiregular module if every element of MM is II-semiregular. RR is called II-semiregular if RR{}_{R}R is an II-semiregular module. Note that II-semiregular rings are left-right symmetric and RR is (δ\delta-) semiregular if and only RR is (δ(RR))\delta(_{R}R)-) J(R)J(R)-semiregular. MM is called an II-semiperfect module [11] if for every submodule KK of MM, there is a decomposition M=ABM=A\oplus B such that AA is projective, AKA\subseteq K and KBIMK\cap B\subseteq IM. RR is called II-semiperfect if RR{}_{R}R is an II-semiperfect module. Note that RR is (δ\delta-) semiperfect if and only RR is (δ(RR))\delta(_{R}R)-) J(R)J(R)-semiperfect. For other standard definitions we refer to [3, 10, 13].

In this note all rings are associative with identity and all modules are unital left modules unless specified otherwise. Let RR be a ring and MM a module. We use Rad(M)Rad(M), Soc(M)Soc(M), Z(M)Z(M) to indicate the Jacobson radical, the socle, the singular submodule of MM respectively. J(R)J(R) is the radical of RR and II is an ideal of RR.

2. II-semiregular(II-semiperfect) rings and projective II-covers

In this section, we introduce the notion of PSD submodules of modules and use this to define projective II-covers which are a generalization of some well-known projective covers. Characterizations of II-semiregular and II-semiperfect rings are given by projective II-covers. We begin this section with the following definitions.

Definition 2.1.

Let II be an ideal of RR and NMN\leq M. NN is PSD in MM if there exists a projective summand SS of MM such that SNS\leq N and M=SXM=S\oplus X whenever N+X=MN+X=M for any submodule XMX\leq M. MM is PSD for II if any submodule of IMIM is PSD in MM. RR is a left PSD ring for II if any finitely generated free left RR-module is PSD for II.

Lemma 2.2.

Let NN be a direct summand of MM and ANA\leq N. Then AA is PSD in MM if and only if AA is PSD in NN.

Proof.

``\Rightarrow" Since NN is a direct summand of MM, M=NKM=N\oplus K for some submodule KMK\leq M. Suppose that N=A+X,XNN=A+X,X\leq N, then M=A+(XK)M=A+(X\oplus K). Since AA is PSD in MM, there is a projective direct summand YY of MM such that YAY\leq A and M=YXKM=Y\oplus X\oplus K, and hence N=NM=XYN=N\cap M=X\oplus Y.

``\Leftarrow" Let M=A+L,LMM=A+L,L\leq M. Then N=NM=A+NLN=N\cap M=A+N\cap L. Since AA is PSD in NN, there is a projective summand KK of NN with KAK\leq A such that N=K(NL)N=K\oplus(N\cap L). It is easy to see that KL=0K\cap L=0. Next we only show that M=K+LM=K+L. Let mMm\in M, then m=a+lm=a+l, aA,lLa\in A,l\in L. Since a=k+s,kK,sNLa=k+s,k\in K,s\in N\cap L, m=k+s+lm=k+s+l. Note that s+lLs+l\in L, so mK+Lm\in K+L, and hence M=K+LM=K+L, as required.

Corollary 2.3.

Let MM be a RR-module. If MM is PSD for an ideal II of RR, then any direct summand of MM is PSD for II.

Corollary 2.4.

A ring RR is a left PSD ring for an ideal II if and only if any finitely generated projective left RR-module is PSD for II.

Proposition 2.5.

Let M=M1M2M=M_{1}\oplus M_{2}. If N1N_{1} is PSD in M1M_{1} and N2N_{2} is PSD in M2M_{2}, then N1N2N_{1}\oplus N_{2} is PSD in MM.

Proof.

Let M=N1N2+L,LMM=N_{1}\oplus N_{2}+L,L\leq M. Since N1N_{1} is PSD in M1M_{1}, N1N_{1} is PSD in MM by Lemma 2.2. Thus there is a projective summand S1S_{1} of MM with S1N1S_{1}\subseteq N_{1} such that M=S1(N2+L)M=S_{1}\oplus(N_{2}+L). Similarly, there exists a projective summand S2S_{2} of MM with S2N2S_{2}\subseteq N_{2} such that M=S1S2+LM=S_{1}\oplus S_{2}+L. The rest is obvious. ∎

Definition 2.6.

A pair (P,f)(P,f) is called a projective II-cover of MM if PP is projective, ff is an epimorphism from PP to MM such that KerfIPKerf\leq IP, and KerfKerf is PSD in PP.

It is easy to see that a module MM has a projective δ(RR)\delta(_{R}R)-cover (projective J(R)J(R)-cover, respectively) if and only if MM has a projective δ\delta-cover (projective cover, respectively) by [1, Proposition 3.6].

Proposition 2.7.

If each fi:PiMi,(i=1,2,,n)f_{i}:P_{i}\rightarrow M_{i},(i=1,2,\cdot\cdot\cdot,n) is a projective II-cover, then i=1nfi:i=1nPii=1nMi\oplus_{i=1}^{n}f_{i}:\oplus_{i=1}^{n}P_{i}\rightarrow\oplus_{i=1}^{n}M_{i} is a projective II-cover.

Proof.

By Proposition 2.5 and the definition of projective II-covers. ∎

Lemma 2.8.

Let II be an ideal of RR and f:PMf:P\rightarrow M a projective II-cover. If QQ is projective and g:QMg:Q\rightarrow M is epic. Then there are decompositions P=ABP=A\oplus B and Q=XYQ=X\oplus Y such that

  1. (1)

    AXA\cong X;

  2. (2)

    fA:AMf\mid_{A}:A\rightarrow M is a projective II-cover;

  3. (3)

    gX:XMg\mid_{X}:X\rightarrow M is a projective II-cover;

  4. (4)

    BKerfB\subseteq Kerf, YKergY\subseteq Kerg.

Proof.

Since QQ is projective, there is a homomorphism h:QPh:Q\rightarrow P such that fh=gfh=g, and so P=h(Q)+KerfP=h(Q)+Kerf. Since KerfKerf is PSD in PP, there is a direct summand BB of PP such that P=ABP=A\oplus B with A=h(Q),BKerfA=h(Q),B\subseteq Kerf. We shall show that fA:AMf\mid_{A}:A\rightarrow M is a projective II-cover. It is clear that KerfA=AKerfAIP=IAKerf\mid_{A}=A\cap Kerf\subseteq A\cap IP=IA. Let KerfA+L=A,LAKerf\mid_{A}+L=A,L\leq A. Then P=KerfA+LB=Kerf+LBP=Kerf\mid_{A}+L\oplus B=Kerf+L\oplus B. Since KerfKerf is PSD in PP, there is a direct summand KK of PP (KKerfK\subseteq Kerf) such that P=LKBP=L\oplus K\oplus B, and hence A=AP=A(LKB)=L(A(KB))A=A\cap P=A\cap(L\oplus K\oplus B)=L\oplus(A\cap(K\oplus B)). It is easy to see A(KB)KerfAA\cap(K\oplus B)\subseteq Kerf\mid_{A}, so KerfAKerf\mid_{A} is PSD in AA. Thus fA:AMf\mid_{A}:A\rightarrow M is a projective II-cover. Since AA is projective, h:QAh:Q\rightarrow A splits, and hence there is a homomorphism q:AQq:A\rightarrow Q such that hq=1Ahq=1_{A}. So Q=XY,X=q(A),Y=KerhQ=X\oplus Y,X=q(A),Y=Kerh. Since X=q(A)X=q(A), AXA\cong X. Next we show that gX:XMg\mid_{X}:X\rightarrow M is a projective II-cover. Since g(X)=fh(X)=fh(X+Y)=fh(Q)=Mg(X)=fh(X)=fh(X+Y)=fh(Q)=M, gX:XMg\mid_{X}:X\rightarrow M is epic. We have KergX=q(KerfA)q(IA)=IXKerg\mid_{X}=q(Kerf\mid_{A})\subseteq q(IA)=IX. Now we only show that KergXKerg\mid_{X} is PSD in XX. Assume that X=KergX+N,XNX=Kerg\mid_{X}+N,X\leq N, then A=h(KergX)+h(N)=KerfA+h(N)A=h(Kerg\mid_{X})+h(N)=Kerf\mid_{A}+h(N). Since KerfAKerf\mid_{A} is PSD in AA, there is a direct summand ZZ of AA (ZKerfA)Z\subseteq Kerf\mid_{A}) such that A=Zh(N)A=Z\oplus h(N). We know that hX:XAh\mid_{X}:X\rightarrow A is isomorphic, so X=hX1(Z)NX=h\mid_{X}^{-1}(Z)\oplus N. It is easy to verify that hX1(Z)KergXh\mid_{X}^{-1}(Z)\subseteq Kerg\mid_{X}, as required. ∎

Lemma 2.9.

Let II be an ideal of a ring RR, MM is a projective RR-module and NMN\leq M. Consider the following conditions:

  1. (1)

    M/NM/N has a projective II-cover.

  2. (2)

    M=YX,YN,XNIMM=Y\oplus X,Y\leq N,X\cap N\leq IM.

Then (1)(2),(2)(1)(1)\Rightarrow(2),(2)\Rightarrow(1) if MM is PSD for II.

Proof.

(1)(2)``(1)\Rightarrow(2)" Let f:PM/Nf:P\rightarrow M/N be a projective II-cover and g:MM/Ng:M\rightarrow M/N be the canonical epimorphism. By Lemma 2.8, M=YX,YKerg=NM=Y\oplus X,Y\subseteq Kerg=N and KergX:XM/NKerg\mid_{X}:X\rightarrow M/N is a projective II-cover, and so KergX=XNIXIMKerg\mid_{X}=X\cap N\subseteq IX\subseteq IM.

(2)(1)``(2)\Rightarrow(1)" Let f:XM/Nf:X\rightarrow M/N with f(x)=x+Nf(x)=x+N. It is easy to see that f:XM/Nf:X\rightarrow M/N is a projective II-cover by Lemma 2.2 and assumptions. ∎

With Lemma 2.9, we can give the following characterization of II-semiregular rings related to projective II-covers.

Theorem 2.10.

Let II be an ideal of RR. Consider the following conditions:

  1. (1)

    Every finitely presented RR-module has a projective II-cover.

  2. (2)

    For every finitely generated left ideal KK of RR, R/KR/K has a projective II-cover.

  3. (3)

    Every cyclically presented left RR-module has a projective II-cover.

  4. (4)

    RR is II-semiregular.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II.

Proof.

By Lemma 2.9, the rest is similar to [1, Theorem 3.11]. ∎

If I=δ(RR)I=\delta{(_{R}R)} or J(R)J{(R)}, then RR is a left PSD ring for II, and hence Theorem 2.10 gives the characterizations of δ\delta-semiregular rings [20] and semiregular rings [10]. Since if RR is Z(RR)Z(_{R}R)-semiregular, then Z(RR)=J(R)δ(RR)Z(_{R}R)=J(R)\subseteq\delta(_{R}R), we have the following result.

Corollary 2.11.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a projective Z(RR)Z(_{R}R)-cover.

Since if ISoc(RR)I\leq Soc(_{R}R), then RR is a left PSD ring for II, and hence we have

Corollary 2.12.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Soc(RR)Soc(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a projective Soc(RR)Soc(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a projective Soc(RR)Soc(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a projective Soc(RR)Soc(_{R}R)-cover.

Next we shall consider the characterizations of II-semiperfect rings.

Theorem 2.13.

Let II be an ideal of RR. Consider the following conditions:

  1. (1)

    Every finitely generated RR-module has a projective II-cover.

  2. (2)

    Every factor module of RR{}_{R}R has a projective II-cover.

  3. (3)

    For every countably generated left ideal KK of RR, R/KR/K has a projective II-cover.

  4. (4)

    RR is II-semiperfect.

  5. (5)

    Every simple RR-module has a projective II-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a projective II-cover.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4) and (1)(5)(6)(1)\Rightarrow(5)\Rightarrow(6), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II; and (6)(4)(6)\Rightarrow(4) if II is PSD in RR{}_{R}R.

Proof.

Similar to [1, Theorem 4.8] by Lemma 2.9. ∎

When I=δ(RR)I=\delta(_{R}R) or Soc(RR)Soc(_{R}R) or J(R)J(R), Theorem 2.13 gives the characterizations of (δ\delta-, Soc(RR)Soc(_{R}R)-) semiperfect rings (See [20], [11], [10]).

Corollary 2.14.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiperfect.

  2. (2)

    Every finitely generated RR-module has a projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a projective Z(RR)Z(_{R}R)-cover.

  5. (5)

    Every simple RR-module has a projective Z(RR)Z(_{R}R)-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a projective Z(RR)Z(_{R}R)-cover.

3. II-semiregular(II-semiperfect) rings and locally projective II-covers

Ding and Chen [5] and Xue [18] gave some characterizations of rings by locally projective covers. Inspired by those, we introduce the notion of locally projective II-covers and use it to characterize II-semiregular and II-semiperfect rings in this section. Firstly, we recall some definitions and facts. A module PP is called locally projective [18, 21] in case it satisfies any of the following equivalent condition: (a)(a) if AA and BB are modules, g:ABg:A\rightarrow B is an epimorphism and f:PBf:P\rightarrow B is a homomorphism then for every finitely generated (cyclic) submodule P0P_{0} of PP there is a homomorphism h:PAh:P\rightarrow A such that fP0=ghP0f\mid_{P_{0}}=gh\mid_{P_{0}}; (b)(b) if MM is a module and f:MPf:M\rightarrow P is an epimorphism then for every finitely generated (cyclic) submodule P0P_{0} of PP there is a homomorphism h:PMh:P\rightarrow M such that fgP0=1P0fg\mid_{P_{0}}=1\mid_{P_{0}}. Clearly, every finitely generated locally projective module is projective. The following facts are well known. (1) a direct sum of locally projective modules is locally projective; (2) any direct summand of a locally projective module is locally projectvie; (3) if PP is a locally projective module, then (1) Rad(P)=J(R)PRad(P)=J(R)P; (2) if Rad(P)=PRad(P)=P, then P=0P=0. We also recall that a pair (P,f)(P,f) is called a locally projective (δ\delta-) cover of MM if PP is locally projective, ff is an epimorphism from PP to MM such that KerfPKerf\ll P (KerfδPKerf\ll_{\delta}P).

Definition 3.1.

A pair (P,f)(P,f) is called a locally projective II-cover of MM if PP is locally projective, ff is an epimorphism from PP to MM such that KerfIPKerf\leq IP, and KerfKerf is PSD in PP.

It is easy to see that a module MM has a locally projective 0-cover if and only if MM is locally projective.

Proposition 3.2.

If each fi:PiMi,(i=1,2,,n)f_{i}:P_{i}\rightarrow M_{i},(i=1,2,\cdot\cdot\cdot,n) is a locally projective II-cover, then i=1nfi:i=1nPii=1nMi\oplus_{i=1}^{n}f_{i}:\oplus_{i=1}^{n}P_{i}\rightarrow\oplus_{i=1}^{n}M_{i} is a locally projective II-cover.

Proposition 3.3.

A module MM has a locally projective J(R)J(R)-cover if and only if MM has a locally projective cover.

Proof.

``\Leftarrow" is clear.

``\Rightarrow" Let f:PMf:P\rightarrow M be a locally projective J(R)J(R)-cover. Then KerfJ(R)PKerf\subseteq J(R)P, KerfKerf is PSD in PP. Next we shall show that KerfPKerf\ll P. Let Kerf+L=P,LMKerf+L=P,L\leq M. Since KerfKerf is PSD in PP, there is a summand KK of PP with KKerfK\leq Kerf such that KL=PK\oplus L=P. Since Rad(K)Rad(L)=Rad(P)Rad(K)\oplus Rad(L)=Rad(P) and KRad(P)K\subseteq Rad(P), Rad(K)=KRad(K)=K. Since KK is locally projective, K=0K=0, and hence L=PL=P, as desired. ∎

The following lemma is a key result of this section.

Lemma 3.4.

Let f:PMf:P\rightarrow M be a locally projective II-cover. If MM is finitely generated, then f:PMf:P\rightarrow M is a projective II-cover.

Proof.

It suffices to prove that PP is projective. Since f:PMf:P\rightarrow M is a locally projective II-cover and MM is finitely generated, there is a finitely generated submodule P0P_{0} of PP such that P0+Kerf=PP_{0}+Kerf=P. Note that KerfKerf is PSD in PP, there is a projective summand KK of PP with KKerfK\subseteq Kerf such that P0K=PP_{0}\oplus K=P. Since PP is locally projective, P0P_{0} is locally projective. Since P0P_{0} is finitely generated, P0P_{0} is projective. Thus PP is projective, as required.

Theorem 3.5.

Let II be an ideal of RR. Consider the following conditions:

  1. (1)

    Every finitely presented RR-module has a locally projective II-cover.

  2. (2)

    For every finitely generated left ideal KK of RR, R/KR/K has a locally projective II-cover.

  3. (3)

    Every cyclically presented left RR-module has a locally projective II-cover.

  4. (4)

    RR is II-semiregular.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II.

Proof.

(1)(2)(3)``(1)\Rightarrow(2)\Rightarrow(3)" are clear.

(3)(4)``(3)\Rightarrow(4)" It follows by Lemma 3.4 and Theorem 2.10.

(4)(1)``(4)\Rightarrow(1)" is clear by Theorem 2.10. ∎

If I=δ(RR)I=\delta{(_{R}R)} or J(R)J{(R)}, then RR is a left PSD ring for II.

Corollary 3.6.

The following conditions are equivalent for a ring RR.

  1. (1)

    RR is semiregular.

  2. (2)

    Every finitely presented RR-module has a locally projective cover.

  3. (3)

    For every finitely generated left ideal KK of RR, R/KR/K has a locally projective cover.

  4. (4)

    Every cyclically presented left RR-module has a locally projective cover.

Corollary 3.7.

The following conditions are equivalent for a ring RR.

  1. (1)

    RR is δ\delta-semiregular.

  2. (2)

    Every finitely presented RR-module has a locally projective δ\delta-cover.

  3. (3)

    For every finitely generated left ideal KK of RR, R/KR/K has a locally projective δ\delta-cover.

  4. (4)

    Every cyclically presented left RR-module has a locally projective δ\delta-cover.

Proof.

It follows by Theorem 2.10 and Lemma 3.4. ∎

Since if RR is Z(RR)Z(_{R}R)-semiregular, then Z(RR)=J(R)δ(RR)Z(_{R}R)=J(R)\subseteq\delta(_{R}R), we have the following result.

Corollary 3.8.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a locally projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a locally projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a locally projective Z(RR)Z(_{R}R)-cover.

Since if ISoc(RR)I\leq Soc(_{R}R), then RR is a left PSD ring for II, and hence we have

Corollary 3.9.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Soc(RR)Soc(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a locally projective Soc(RR)Soc(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a locally projective Soc(RR)Soc(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a locally projective Soc(RR)Soc(_{R}R)-cover.

Next we shall consider the characterizations of II-semiperfect rings.

Theorem 3.10.

Let II be an ideal of RR. Consider the following conditions:

  1. (1)

    Every finitely generated RR-module has a locally projective II-cover.

  2. (2)

    Every factor module of RR{}_{R}R has a locally projective II-cover.

  3. (3)

    For every countably generated left ideal KK of RR, R/KR/K has a locally projective II-cover.

  4. (4)

    RR is II-semiperfect.

  5. (5)

    Every simple RR-module has a locally projective II-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a locally projective II-cover.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4) and (1)(5)(6)(1)\Rightarrow(5)\Rightarrow(6), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II; and (6)(4)(6)\Rightarrow(4) if II is PSD in RR{}_{R}R.

Proof.

It follows by Theorem 2.13 and Lemma 3.4. ∎

When I=δ(RR)I=\delta(_{R}R) or J(R)J(R), we have the following.

Corollary 3.11.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is semiperfect.

  2. (2)

    Every finitely generated RR-module has a locally projective cover.

  3. (3)

    Every factor module of RR{}_{R}R has a locally projective cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a locally projective cover.

  5. (5)

    Every simple RR-module has a locally projective cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a locally projective cover.

Corollary 3.12.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is δ\delta-semiperfect.

  2. (2)

    Every finitely generated RR-module has a locally projective δ\delta-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a locally projective δ\delta-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a locally projective δ\delta-cover.

  5. (5)

    Every simple RR-module has a locally projective δ\delta-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a locally projective δ\delta-cover.

Proof.

By Lemma 3.4 and Theorem 3.10. ∎

Corollary 3.13.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiperfect.

  2. (2)

    Every finitely generated RR-module has a locally projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a locally projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a locally projective Z(RR)Z(_{R}R)-cover.

  5. (5)

    Every simple RR-module has a locally projective Z(RR)Z(_{R}R)-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a locally projective Z(RR)Z(_{R}R)-cover.

Corollary 3.14.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is Soc(RR)Soc(_{R}R)-semiperfect.

  2. (2)

    Every finitely generated RR-module has a locally projective Soc(RR)Soc(_{R}R)-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a locally projective Soc(RR)Soc(_{R}R)-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a locally projective Soc(RR)Soc(_{R}R)-cover.

  5. (5)

    Every simple RR-module has a locally projective Soc(RR)Soc(_{R}R)-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a locally projective Soc(RR)Soc(_{R}R)-cover.

4. II-semiregular(II-semiperfect) rings characterized by projectivity classes

Wang [14] gave characterizations of semiregular rings and semiperfect rings by introducing the concept of projectivity classes of modules. Motivated by this, we shall characterize II-semiregular rings and II-semiperfect rings in term of projectivity classes of modules in this section.

Definition 4.1.

( [14] ) A class 𝒫\mathscr{P} of RR-modules is called a projectivity class if it contains all self-projective modules and for every module MM and every projective module PP with an epimorphism f:PMf:P\rightarrow M, PM𝒫P\oplus M\in\mathscr{P} implies that MM is projective.

Example 4.2.

1. The class of all quasi-projective modules is a projectivity class.

2. The class of all weakly quasi-projective modules in the sense of Rangaswamy and Vanaja is a projectivity class.

3. The class of all pseudo-projective modules is a projectivity class.

4. The class of all direct-projective modules is a projectivity class.

5. For any perfect ring RR, the class of all discrete RR-modules is a projectivity class.

For a projectivity class 𝒫\mathscr{P}, we introduce the following concept.

Definition 4.3.

Let II be an ideal of a ring RR and MM be a module. We call an epimorphism f:PMf:P\rightarrow M a 𝒫\mathscr{P}-projective II-cover of MM if P𝒫P\in\mathscr{P} and KerfIPKerf\leq IP, KerfKerf is PSD in PP.

Lemma 4.4.

Let II be an ideal of a ring RR, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Suppose that PP is a projective module and there is an epimorphism f:PMf:P\rightarrow M. If PMP\oplus M has a 𝒫\mathscr{P}-projective II-cover, then MM has a projective II-cover.

Proof.

Let g:QPMg:Q\rightarrow P\oplus M be a 𝒫\mathscr{P}-projective II-cover of PMP\oplus M. We have an exact sequence 0g1(M)QϕgP00\rightarrow g^{-1}(M)\rightarrow Q\stackrel{{\scriptstyle\phi g}}{{\rightarrow}}P\rightarrow 0, where ϕ:PMP\phi:P\oplus M\rightarrow P is the projection map. Since PP is projective, QPg1(M)Q\cong P\oplus g^{-1}(M) and Ker(ϕg)=g1(M)(\phi g)=g^{-1}(M) is a direct summand of QQ. Note that Kerg is PSD in QQ, and so Kerg is PSD in g1(M)g^{-1}(M) by Lemma 2.2. It is easy to see that KergIg1(M)g\subseteq Ig^{-1}(M). Clearly, we have an exact sequence 00\rightarrowKergg1(M)gM0g\rightarrow g^{-1}(M)\stackrel{{\scriptstyle g}}{{\rightarrow}}M\rightarrow 0. So it suffices to show that g1(M)g^{-1}(M) is projective. Since PP is projective with an epimorphism f:PMf:P\rightarrow M, and g:g1(M)Mg:g^{-1}(M)\rightarrow M is an epimorphism, there is a homomorphism α:Pg1(M)\alpha:P\rightarrow g^{-1}(M) such that gα=fg\alpha=f, and hence Imα+\alpha+Kerg=g1(M)g=g^{-1}(M). Since Kergg is PSD in g1(M)g^{-1}(M), there is a projective submodule LL of Kergg such that ImαL=g1(M)\alpha\oplus L=g^{-1}(M). Thus QPQ\cong P\oplusImαL\alpha\oplus L belongs to 𝒫\mathscr{P}. Since 𝒫\mathscr{P} is closed under taking direct summands, PP\oplusImα𝒫\alpha\in\mathscr{P}, and there is an epimorphism PP\rightarrowImα\alpha, Imα\alpha is projective. Thus g1(M)=g^{-1}(M)=ImαL\alpha\oplus L is projective, as desired. ∎

Corollary 4.5.

Let II be an ideal of a ring RR, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Consider the following statements:

  1. (1)

    Every finitely presented RR-module has a 𝒫\mathscr{P}-projective II-cover.

  2. (2)

    For every finitely generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective II-cover.

  3. (3)

    Every cyclically presented left RR-module has a 𝒫\mathscr{P}-projective II-cover.

  4. (4)

    RR is II-semiregular.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II.

Proof.

(1)(2)(3)``(1)\Rightarrow(2)\Rightarrow(3)" are clear.

(3)(4)``(3)\Rightarrow(4)" RR/RrR\oplus R/Rr has a 𝒫\mathscr{P}-projective by assumption, and hence R/RrR/Rr has a projective II-cover by Lemma 4.4. Thus RR is II-semiregular by Theorem 2.10. (4)(1)``(4)\Rightarrow(1)" is clear by Theorem 2.10. ∎

Corollary 4.6.

Let RR be a ring, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Then the following statements are equivalent.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover..

Corollary 4.7.

Let RR be a ring, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Then the following statements are equivalent.

  1. (1)

    RR is Soc(RR)Soc(_{R}R)-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  3. (3)

    Every finitely presented RR-module has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

Corollary 4.8.

Let 𝒫\mathscr{P} be a projectivity class and which is closed under taking direct summands. The following statements are equivalent for a ring RR:

  1. (1)

    RR is δ\delta-semiregular.

  2. (2)

    Every cyclically presented left RR-module has a 𝒫\mathscr{P}-projective δ\delta-cover.

  3. (3)

    Every finitely presented RR-module has a 𝒫\mathscr{P}-projective δ\delta-cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective δ\delta-cover.

When 𝒫\mathscr{P} is the class of all direct-projective modules, Corollary 4.8 gives [15, Proposition 4.4].

Corollary 4.9.

Let 𝒫\mathscr{P} be a projectivity class and which is closed under taking direct summands. The following statements are equivalent for a ring RR:

  1. (1)

    RR is semiregular.

  2. (2)

    Every cyclically presented left RR-module has a 𝒫\mathscr{P}-projective cover.

  3. (3)

    Every finitely presented RR-module has a 𝒫\mathscr{P}-projective cover.

  4. (4)

    For every finitely generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective cover.

When 𝒫\mathscr{P} is the class of all direct-projective modules, Corollary 4.9 gives [17, Corollary 3.4].

Theorem 4.10.

Let II be an ideal of a ring RR, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Consider the following statements:

  1. (1)

    Every finitely generated RR-module has a 𝒫\mathscr{P}-projective II-cover.

  2. (2)

    Every factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective II-cover.

  3. (3)

    For every countably generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective II-cover.

  4. (4)

    RR is II-semiperfect.

  5. (5)

    Every simple RR-module has a 𝒫\mathscr{P}-projective II-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective II-cover.

Then (1)(2)(3)(4)(1)\Rightarrow(2)\Rightarrow(3)\Rightarrow(4) and (1)(5)(6)(1)\Rightarrow(5)\Rightarrow(6), (4)(1)(4)\Rightarrow(1) if RR is a left PSD ring for II; and (6)(4)(6)\Rightarrow(4) if II is PSD in RR{}_{R}R.

Proof.

Following by Theorem 2.13 and Lemma 3.4. ∎

Corollary 4.11.

Let II be an ideal of a ring RR, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Then the following statements are equivalent.

  1. (1)

    RR is Z(RR)Z(_{R}R)-semiperfect.

  2. (2)

    Every finitely generated RR-module has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  5. (5)

    Every simple RR-module has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective Z(RR)Z(_{R}R)-cover.

Corollary 4.12.

Let II be an ideal of a ring RR, 𝒫\mathscr{P} a projectivity class and which is closed under taking direct summands. Then the following statements are equivalent.

  1. (1)

    RR is Soc(RR)Soc(_{R}R)-semiperfect.

  2. (2)

    Every finitely generated RR-module has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  5. (5)

    Every simple RR-module has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective Soc(RR)Soc(_{R}R)-cover.

Corollary 4.13.

Let 𝒫\mathscr{P} be any projectivity class and which is closed under taking direct summands. The following statements are equivalent for a ring RR.

  1. (1)

    RR is a left δ\delta-semiperfect ring.

  2. (2)

    Every finitely generated RR-module has a 𝒫\mathscr{P}-projective δ\delta-cover.

  3. (3)

    Every factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective δ\delta-cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective δ\delta-cover.

  5. (5)

    Every simple RR-module has a 𝒫\mathscr{P}-projective δ\delta-cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective δ\delta-cover.

Corollary 4.14.

Let 𝒫\mathscr{P} be any projectivity class and which is closed under taking direct summands. The following statements are equivalent for a ring RR.

  1. (1)

    RR is a left semiperfect ring.

  2. (2)

    Every finitely generated RR-module has a 𝒫\mathscr{P}-projective cover.

  3. (3)

    Every factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective cover.

  4. (4)

    For every countably generated left ideal KK of RR, R/KR/K has a 𝒫\mathscr{P}-projective cover.

  5. (5)

    Every simple RR-module has a 𝒫\mathscr{P}-projective cover.

  6. (6)

    Every simple factor module of RR{}_{R}R has a 𝒫\mathscr{P}-projective cover.

5. II-semiregular(II-semiperfect) rings and II-supplemented modules

It is well know that a ring RR is semiperfect if and only if RRR_{R} is a supplemented module if and only if RR{}_{R}R is a supplemented module. We also know that a ring RR is semiregular if and only if RRR_{R} is a finitely supplemented module if and only if RR{}_{R}R is a finitely supplemented module. Here we introduce the notion of II-supplemented modules and use it to characterize II-semiregular(semiperfect) rings.

Let RR be a ring II an ideal of RR, MM a module and N,LMN,L\leq M. NN is called a supplement of LL in MM if N+L=MN+L=M and NN is minimal with respect to this property. Equivalently, M=N+LM=N+L and NLNN\cap L\ll N. A module MM is called supplemented if every submodule of MM has a supplement in MM. NN is called a δ\delta-supplement [6] of LL if M=N+LM=N+L and NLδNN\cap L\ll_{\delta}N. MM is called a δ\delta-supplemented module if every submodule of MM has a δ\delta-supplement. A module MM is said to be δ\delta-lifting [6] if for any submodule NN of MM, there exists a direct summand KK of MM such that KNK\leq N and N/KδM/KN/K\ll_{\delta}M/K, equivalently, for every submodule NN of MM, MM has a decomposition with M=M1M2M=M_{1}\oplus M_{2}, M1NM_{1}\leq N and M2NM_{2}\cap N is δ\delta-small in M2M_{2}. NN is DM in MM [1] if there is a summand SS of MM such that SNS\leq N and M=S+XM=S+X, whenever N+X=MN+X=M for a submodule XX of MM. MM is DM for II if any submodule of IMIM is DM in MM. RR is a left DM ring for II if for any finitely generated free left RR-module is DM for II.

Definition 5.1.

Let RR be a ring and II an ideal of RR, MM a module. MM is called an II-supplemented module (finitely II-supplemented module) if for every submodule (finitely generated submodule ) XX of MM, there is a projective submodule YY of MM such that X+Y=MX+Y=M, XYIYX\cap Y\subseteq IY and XYX\cap Y is DM in YY.

Theorem 5.2.

Let RR be a ring. The following statements are equivalent for a projective module MM.

  1. (1)

    MM is a J(R)J(R)-supplemented module (a δ(RR)\delta(_{R}R)-supplemented module, respectively).

  2. (2)

    MM is a supplemented module ( a δ\delta-supplemented module, respectively).

Proof.

(1)(2)``(1)\Rightarrow(2)" Let MM be a J(R)J(R)-supplemented module (a δ(RR)\delta(_{R}R)-supplemented module, respectively). Then for every submodule XX of MM, there is a projective submodule YY of MM such that X+Y=MX+Y=M , XYJ(R)YX\cap Y\subseteq J(R)Y (XYδ(RR)YX\cap Y\subseteq\delta(_{R}R)Y) and XYX\cap Y is DM in YY. Next we shall show that XYYX\cap Y\ll Y (XYδYX\cap Y\ll_{\delta}Y). Assume that XY+L=Y,LYX\cap Y+L=Y,L\leq Y. Note that XYX\cap Y is DM in YY, there is a summand KK of YY which is contained in XYX\cap Y such that K+L=YK+L=Y. Write Y=KK,KMY=K\oplus K^{\prime},K^{\prime}\leq M, then Rad(K)Rad(K)=Rad(Y)Rad(K)\oplus Rad(K^{\prime})=Rad(Y) (δ(K)δ(K)=δ(Y)\delta(K)\oplus\delta(K^{\prime})=\delta(Y)). Since KRad(Y)K\subseteq Rad(Y) (Kδ(Y)K\subseteq\delta(Y)), K=Rad(K)K=Rad(K) (K=δ(K)K=\delta(K)). Since KK is projective, K=0K=0 (KK is semisimple, thus XYδYX\cap Y\ll_{\delta}Y by [20, Lemma 1.2]), and hence L=YL=Y, so XYYX\cap Y\ll Y.

(2)(1)``(2)\Rightarrow(1)" Let MM be a supplemented module. Then for every submodule XX of MM, there a submodule YY of MM such that X+Y=MX+Y=M and XYYX\cap Y\ll Y. Since MM is projective, YY is a direct summand of MM, and hence YY is projective. It is clear that XYRad(Y)=J(R)YX\cap Y\subseteq Rad(Y)=J(R)Y and XYX\cap Y is DM in YY. (Let MM be a δ\delta-supplemented module. Since MM is projective, MM is δ\delta-lifting. Thus for every submodule XX of MM, there is a direct summand YY of MM such that M=X+YM=X+Y and XYδYX\cap Y\ll_{\delta}Y. The rest is obvious.)

Theorem 5.3.

Let RR be a ring and II an ideal of RR, MM a projective module. Consider the following conditions:

  1. (1)

    MM is an II-supplemented module.

  2. (2)

    MM is an II-semiperfect module.

Then (1)(2)(1)\Rightarrow(2), and (2)(1)(2)\Rightarrow(1) if MM is DM for II.

Proof.

(1)(2)``(1)\Rightarrow(2)" Let MM be an II-supplemented module, then for every submodule XX of MM, there is a projective submodule YY of MM such that M=X+YM=X+Y, XYIYX\cap Y\subseteq IY and XYX\cap Y is DM in YY. We define f:YM/Xf:Y\rightarrow M/X be such that f(y)=y+Xf(y)=y+X. Then ff is an epimorphism with Kerf=XYKerf=X\cap Y, and hence YY is a projective II-cover (in the sense of [1]) of M/XM/X. The rest is obvious by [1, Lemma 3.10].

(2)(1)``(2)\Rightarrow(1)" Let MM be an II-semiperfect module, then for every submodule XX of MM, there is a decomposition M=AYM=A\oplus Y such that AA is projective, AXA\subseteq X and XYIMX\cap Y\subseteq IM. Thus M=X+YM=X+Y, YY is projective, XYIYX\cap Y\subseteq IY. Since MM is DM for II, XYX\cap Y is DM in YY by [1, Lemma 3.2], as desired.

Corollary 5.4.

Let MM be a projective module with Rad(M)MRad(M)\ll M (δ(M)δM\delta(M)\ll_{\delta}M). Then MM is a (δ\delta-)supplemented module if and only if MM is a (δ\delta-)semiperfect module.

Corollary 5.5.

Let RR be a left DM ring and II an ideal of RR. Then RR is an II-semiperfect ring if and only if RR{}_{R}R is an II-supplemented module.

Write I=J(R)I=J(R) or δ(RR)\delta(_{R}R) in Corollary 5.5, since RR is a left DM ring, we have the following.

Corollary 5.6.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is (δ\delta-) semiperfect;

  2. (2)

    RR{}_{R}R is a (δ\delta-) supplemented module;

  3. (3)

    RRR_{R} is a (δ\delta-) supplemented module;

  4. (4)

    RR{}_{R}R is a J(R)J(R)(δ(RR)\delta(_{R}R))-supplemented module;

  5. (5)

    RRR_{R} is a J(R)J(R)(δ(RR)\delta(_{R}R))-supplemented module.

Proof.

It follows by Theorem 5.2 and 5.3. ∎

Similarly, we obtain the following results.

Theorem 5.7.

Let RR be a left DM ring and II an ideal of RR. Then RR is an II-semiregular ring if and only if RR{}_{R}R is a finitely II-supplemented module if and only if RRR_{R} is a finitely II-supplemented module.

Corollary 5.8.

The following statements are equivalent for a ring RR.

  1. (1)

    RR is (δ\delta-) semiregular.

  2. (2)

    RR{}_{R}R is a finite (δ\delta-) supplemented module.

  3. (3)

    RRR_{R} is a finite (δ\delta-) supplemented module.

  4. (4)

    RR{}_{R}R is a finite J(R)J(R)(δ(RR)\delta(_{R}R))-supplemented module.

  5. (5)

    RRR_{R} is a finite J(R)J(R)(δ(RR)\delta(_{R}R))-supplemented module.

Since if ISoc(RR)I\subseteq Soc(_{R}R), then RR is a left DM ring, we have

Corollary 5.9.

A ring RR is Soc(RR)Soc(_{R}R)-semiperfect if and only if RR{}_{R}R is a Soc(RR)Soc(_{R}R)-supplemented module.

Corollary 5.10.

A ring RR is Soc(RR)Soc(_{R}R)-semiregular if and only if RR{}_{R}R is a finitely Soc(RR)Soc(_{R}R)-supplemented module if and only if RRR_{R} is a finitely Soc(RR)Soc(R_{R})-supplemented module.

References

  • [1] M. Alkan, W. K. Nicholson, A. C. Ozcan, A generalization of projective covers, J. Algebra 319 (2008) 4947-4960.
  • [2] M. Alkan, A. C. Ozcan, Semiregular modules with respect to a fully invariant submodule, Comm. Algebra 32(11) (2004) 4285-4301.
  • [3] F. W. Anderson and K. R. Fuller, Rings and Categories of Modules, Springer-Verlag, New York, 1974.
  • [4] J. Clark, C. Lomp, N. Vanaja and R. Wisbauer, Lifting modules, IN:Supplements and Projectivity in Module Theory Sereis: Frontiers in Mathematics, 2006.
  • [5] N. Q. Ding and J. L. Chen, On a Characterization of Perfect Rings, Comm. Algebra, 27 (1999) 785-791.
  • [6] M. T. Kosan, δ\delta-lifting and δ\delta-supplemented modules, Algebra Colloq. 14 (2007) 53-60.
  • [7] S. H. Mohamed and B. J. Müller, Continuous and Discrete Modules, London Math. Soc.; LNS147, Cambridge Univ. Press: Cambridge, 1990.
  • [8] W. K. Nicholson, Semiregular modules and rings, Can. J. Math. 5 (1976) 1105-1120.
  • [9] W. K. Nicholson, Strong lifting, J. Algebra 285 (2005) 795-818.
  • [10] W. K. Nicholson, M. F. Yousif, Quasi-Frobenius Rings, Cambridge Tracts in Math., vol.158, Cambridge University Press, Cambridge, 2003.
  • [11] A. C. Ozcan, M. Alkan, Semiperfect modules with respect to a preradical, Comm. Algebra 34 (2006) 841-856.
  • [12] K. M. Rangaswamy and N. Vanaja, Quasi projectives in Abelian and module categories, Pacific J. math 43(1) (1972) 221-238.
  • [13] A. Tuganbaev, Rings Close to Regular, Kluwer Academic Publishers, Dordrecht, 2002.
  • [14] D. G. Wang, Rings characterized by pojectivity classes, Comm. Algebra 25(1) (1997) 105-116.
  • [15] Y. D. Wang, Relatively semiregular modules, Algebra Colloq. 13 (2006) 647-654.
  • [16] R. Wisbauer, Foundations of Module and Ring Theory, Gordon and Breach, Philadelphia, 1991.
  • [17] W. M. Xue, Semiregular modules and F-semiperfect modules, Comm. Algebra 23(3) (1995) 1035-1046.
  • [18] W. M. Xue, Characterizations of Semiperfect and Perfect Rings, Publ. Mat. 40 (1996) 115-125.
  • [19] M. F. Yousif, Y. Q. Zhou, Semiregular,semiperfect and perfect rings relative to an ideal, Rocky Mountain J. Math. 32 (2002) 1651-1671.
  • [20] Y. Q. Zhou, Generalizations of perfect, semiperfect, and semiregular rings, Algebra Colloq. 3 (2000) 305-318.
  • [21] B. Zimmermann-Huisgen, Pure submodules of direct products of free modules, Math. Ann. 224 (1976) 233-245.
  • [22] B. Zimmermann-Huisgen, Direct products of modules and algebraic compactness, Habilitationsschrift, Tech. Univ. Munich, 1980.