Аннотация:Insurance is the oldest domain of applied probability. Moreover, the mathematical models arising there can be used in other areas of applied probability. So, optimization of insurance models performance and their asymptotic analysis are very important. Modern period in actuarial sciences is characterized by investigation of complex systems and employment of sophisticated mathematical tools. Discrete-time models became popular, since in many cases they describe more precisely the real situation. Hence, we study two models (discrete-time one and continuous-timeone) in the framework of cost approach. Reinsurance, dividends and bank loans are controls in optimization problems. Models stability with respect to small perturbations of underlying distributions is treated as well using the probability metrics.