Feed Forward Enabled Interface of Artificial Neural Network to System Dynamics Modeling for Developing Experiential Expert System for Decentralized Marketing Logistics

Abstract

Neural networks have the potential of accurately describing the behavior of extremely complex systems such as being encountered in decentralized logistics. Learned feed-forward is a great way to reduce dynamic tracking error of a feedback based decentralized logistics control system in those cases where logistics events recur. The present paper primarily proposes a possible interface of ANN into SD modeling offering a tremendous opportunity of developing an Experiential Expert System for logistics control, policy formulation and decision-making. Additionally, in the area of soft computational research targeting optimization solutions, ANN-SD-Genetic Algorithmic integration can also prove to be a potent framework for the optimization of logistical processes and decision variables. Hence, the research proposition suggested here on the whole highlights the prospective systemic interface and compatibility between System Dynamics, Artificial neural network and Genetic Algorithm.