Solving the asset and liability management problem under uncertainty: a simheuristic multiobjective approach

Asset and liability management (ALM) deals with mapping assets to liabilities to maximize profits over time. Under realistic assumptions, solving this optimization problem is critical for insurance companies and banks. Within the context of uncertainty, there is a conflict between the cost of the mapping plan and the risk of default. In this work, we propose a multiobjective algorithm to solve the ALM problem by simultaneously considering both the cost and risk of alternative policies. Initially, a multiobjective greedy heuristic is presented, which is later transformed into a probabilistic algorithm using biased-randomization techniques. To handle uncertainty, the biased-randomized algorithm is further developed into a simheuristic algorithm, integrating Monte Carlo simulation. The results from our computational experiments show that the solutions provided by our simheuristic method outperform the deterministic solutions when these are considered in a realistic scenario under uncertainty.