A Self-Adaptive Agent-Based Dynamic Processes Simulation Modelling Framework

Abstract

Traditional simulation modelling of dynamic processes aimed at domain-specific simulation with specific simulation objectives. Studies of recent years uncovered drawbacks of domain-specific simulation models for dynamic processes which include unreliable simulation results and questionable robustness of simulation model. Generic agent-based simulation modelling frameworks tried to bridge the gap being a platform to simulate different domain-specific simulation models. However, the efforts were not continued, most probably due to the main gap of having domain-specific models developed for each specific simulation objectives were not addressed. Thus, the aim of this research to bridge the gap by having a self-adaptive simulation model. Dynamic processes from three different domains. For instance, transportation, education and security were used as benchmark of self-capability of the simulation model. Key parameters, self-adaptive algorithms were formulated and implemented through simulation model. The self-adaptive simulation model was able to construct simulation model and generate replicable results based on different combinations of key parameter inputs entered into the simulation system. This achievement opened the possibility of having a self-adaptive simulation modelling of dynamic processes for three different domains.