Page 292 - Data Science Algorithms in a Week

Agent-Based Modeling Simulation and Its Application to Ecommerce 273

DISCUSSION AND FUTURE DIRECTION

The developed simulation models serve as a testbed for managing control actions by
incorporating fluctuations and stochasticity. The system dynamics model captures a high
level abstraction of the system. A multi-model paradigm consisting of agent based
simulation allows appropriate choice of techniques that take into consideration different
components of the system.
In online consumer-to-consumer lending, risks and uncertainties pervade aspects of
operation. The model uses consumers’ historical payments, outstanding debts, amount of
credit available, income and length of credit history to make its calculations. The
framework offers a structured approach that incorporates business processes and
stakeholder requirements and lends its use to ecommerce systems.
The developed simulation model takes into consideration difference in customer
characteristics and stochasticity in demand patterns. The framework provides insights to
the overall behavior of consumer-to-consumer ecommerce complex systems. This in turn
provides insights on the profitability of the business model and strategies for improving
system performance. The result is a recommendation for a course of action which
complements management’s expertise and intuition.
An extension to this study will be to explore the case where a borrower’s request is
met by multiple lenders, and how such strategy impacts individual and system
performance. There is also room to improve on the risk classification phase. Validity of
the results hinges on correct interpretation of the output of the model. As a result, there is
a need to also improve the accuracy of neural network prediction algorithm to encompass
a system that perpetually improves based on learning. Further research can also
investigate to what extent P2P models can reduce costs and fees and if such reduction is
worth the associated risk.
It is expected that conceptual modeling approaches will continue to be a beneficial
approach for analyzing consumer-to-consumer complex systems. This study lays a
foundation for future research to expand on the guidelines and simulation development in
modeling the operations of an organization.

REFERENCES

An, Lianjun, & Jeng, J.-J. (2005). On developing system dynamics model for business
process simulation. In Simulation Conference, 2005 Proceedings of the Winter (p. 10
pp.-). https://doi.org/10.1109/WSC.2005.1574489.
An, Liping, Du, Y., & Tong, L. (2016). Study on Return Policy in E-Commerce
Environment Based on System Dynamics. In Proceedings of the 2nd Information