\(\Pi\)-coherent rings and FGT-injective dimension (Q2564839): Difference between revisions

Let \(R\) be a ring with identity. A right \(R\)-module \(M\) is called FGT-injective if \(\text{Ext}^1_R(A,M)=0\) for every finitely generated torsionless right \(R\)-module \(A\). This type of injectivity gives an FGT-injective dimension of \(M\) and a global dimension for \(R\), denoted by \(\text{rFGT-I.dim }R\). A right \(R\)-module \(M\) is called a \(W\)-module if \(M\cong A^*/K\), where \(A\) is a finitely generated left \(R\)-module and \(A^*=\text{Hom}_R(M,R)\). Then a right \(R\)-module \(M\) is called \(\Pi\)-coherent if \(M\) and all of its \(W\)-submodules are finitely generated. Also, \(R\) is called \(\Pi\)-coherent if every finitely generated submodule of a direct product of copies of \(R_R\) is finitely presented. Some basic properties of \(\Pi\)-coherent rings are given. For example, \(\Pi\)-coherence is a Morita equivalent property, and the following statements are equivalent for a \(\Pi\)-coherent ring \(R\): (1) \(R\) is right FGT-self-injective, (2) every finitely generated torsionless left \(R\)-module is reflexive, (3) every flat right \(R\)-module is FGT-injective, and (4) every projective right \(R\)-module is FGT-injective. Then the FGT-injective dimension is discussed for \(\Pi\)-coherent rings. For a left and right \(\Pi\)-coherent ring, the following statements hold: (1) \(\text{FGT-inj dim}_RR\leq 1\) if and only if every finitely generated torsionless module is reflexive, and (2) \(\text{rFGT-I.dim }R\leq\text{w.gl.dim }R\leq\text{lFGT-I.dim }R+1\leq\text{l.gl.dim }R\). Several specialized results are given for semilocal \(\Pi\)-coherent rings.