Modeling Decision Strategies in Supply-Chains - An Agent-based Approach for Supply Network Planning in Dynamic Markets

Beschreibung

In global markets, one major issue for manufacturing organizations is to adopt the supply-chain to the dynamics of consumer markets. Heterogeneous market structures, trends, customer needs and product seasons are just a few of the challenges a global supply-chain needs to deal with. Future IT systems need to be able to respond to stochastic and even unpredictable market changes in a better and more flexible way than classical enterprise resource planning (ERP) or supply-chain management (SCM) systems.

This study is focusing on the optimization of supply networks for highly volatile consumer markets. Both, the market dynamic and the strategies for supply network planning are modeled using an agent-based approach, which captures the dynamic of the decision problem implicitly. The two agent-based models are combined to provide a dynamic formulation of the decision problem for further optimization tasks. The interaction of the two agent-based models, one for the consumer market and one for the supply network optimization is studied and selected stylized facts reproduced.

The market model describes the dynamic of consumer decisions influenced by classical marketing variables like price, product features, advertising, life-cycle, and trend. The purchase decision of the consumer agents depends on their preferences which change over time. The market dynamics is generating the needs for the supply network. The optimal decisions for the supply network are generated by the interaction and collaboration of the agents. The resulting dynamic network structure is influencing the price and therefor the market dynamics. In this study the optimal marketing mix and network structure will be identified for different dynamic market scenarios. The result will be compared to classical linear programming solutions.

Zeitraum	29 Apr. 2011 → 2 Mai 2011
Ereignistitel	22nd Annual POMS Conference
Veranstaltungstyp	Keine Angaben
Bekanntheitsgrad	International