Computational optimizations of nested simulations utilized for decision-making support

Abstract

Nested simulations present a general method suitable for use in realizing a multi-trajectory simulation or as a decision support in a simulator. The principle of nested simulation (as a decision support) is to find a solution to a problem using other time-limited simulations which verify alternative options. After the nested simulations have finished, the solutions of individual alternatives are assessed and the best solution is applied to the main simulation. The aim of this article is to test various exact and heuristic approaches to reduce the computational complexity of nested multi-simulations. The basic method uses the recorded states from previous simulations to avoid duplicate calculations leading to the same states. Other measures are heuristic and are based on the detection of similar states/replications during the calculation.