regularization at constitutive model level

Hi

I am a post doc working on elasto viscoplastic constitutivemodels to simulate creep and time dependant behaviour in soft clays. The models,which are employed as user defined constitutive models for FE simulations ofgeotechnical applications, are developed based on Perzyna's over stress theory. ButI am experiencing difficulties when modelling problems close to failure. Inorder to overcome this,  I was told todevelop a non-local formulation of the models. However, regularization atconstitutive model level is new for me. I would really appreciate if peoplehere could introduce me sources or literatures where I can find detailedexplanation of such non-local techniques (particularly applicable to soil viscoplasticconstitutive models).

very many thanks

reza