Family Financial Planning based on Probabilistic Planners

Abstract

Financial planning is a step-by-step guiding approach to meet financial goals in life, such as controlling income, expenses, and investments to manage money and estate, which is beneficial for both organizations and families. The functionality of family financial planning is reflected in its charges, which motivated automated family financial planning both from scholars and finance practitioners. However, previous papers and tools rely excessively on linear programming, are hard to explain and interpret, and are incompatible with probabilistic and paralleled natures of financial actions. In the meantime, a recently developed subarea of planning based on the Relational Dynamic Influence Diagram Language (RDDL) happened to model probabilistic and paralleled natures of actions. I.e., probabilistic planning, whose models are explainable and transparent but with few applications. Therefore, this paper proposed a simulation-based approach called the F2P3 (Family Financial Planning based on Probabilistic Planner), whose first action is collecting family information through interactions. Then, the F2P3 models the family financial planning as a probabilistic planning problem and translates it into an RDDL file. At last, it runs probabilistic planners with the RDDL file as a new paradigm of the planning application. Experiments showed that the F2P3 could model the probabilistic and paralleled natures of financial actions and be more comprehensive than existing academic methods and cost-effective than business approaches.