Catastrophic dependent systemic losses have analytically intractable multidimensional probability distributions dependent on exogenous shocks, interactions among goals and constraints of the involved actors and systems, activities of economic sectors, structural and environmental standards, critical infrastructure in place, feasible mitigation and adaptation structural and financial measures, investment potentials, etc. For the analysis of the systemic risks we argue for the design of proper Decision Support Systems (DSSs) and integrated catastrophe analysis and management modeling approaches similar to ISCRiMM model of IIASA. We discuss several important aspects and components of the ISCRiMM. This includes considerations of systemic risks, safety and security constraints, the necessity of robust ex-ante loss reduction and ex-post emergency response and BBB actions, structural and financial measures, the need for stochastic catastrophe models (scenario generators), and proper stochastic optimization solution procedures to enable the decision-support regarding coherent systemic ex-ante and ex-post preventing and coping solutions for dealing with catastrophes. The diversion of capital from ex-post measures to ex-ante investments into structural loss reduction measures can essentially reduce the dependencies among losses and, hence, decrease overall vulnerability, stabilize insurance mechanisms, reduce the demand for ex-post risk sharing and restoration efforts. One of the ISCRiMM submodels is Vulnerability assessment model. In the paper we discuss different methodologies and models for vulnerabilities analysis and modeling, and how they can be effectively integrated within ISCRiMM. In particular, vulnerability models can be based on AI, statistical and machine learning principles, which provide an effective means of incorporating them into ISCRiMM and designing optimal and robust interdependent ex-ate and ex-post measures decreasing vulnerabilities and increasing resilience and BBB capacities.