Register now After registration you will be able to apply for this opportunity online.
When do supply chains between selfish agents fail?
In this project, we will investigate when global supply chains fail and how the productivity level is impacted by local interaction structures and selfish decision-making dynamics via mathematical analysis and simulation. We will extend existing models on the two-player supply chain game to multi-player supply chains with general connectivity structures modeled via graph theory, and investigate various player dynamics (e.g. consensus, best response, gradient descent) in combination with different interconnection structures(e.g. trees, small-world network, star) to study the stability of the overall supply chain.
As modern technology digitizes sectors including agriculture, power grids, and transportation networks, a network of interconnected and self-interested decision-makers emerges as a prominent feature of urban supply chains. Whereas traditional supply chains tend to follow a multi-tier, tree-like structure, e.g. producers supply distributors, who supply retailers, who supply consumers. Increased global mobility and the formation of retail giants have introduced greater complexity in supply chain pathways. For example, a local bakery may supply a grocery store chain with baked goods, while the grocery store chain may supply the bakery with flour and other baking necessities. How do these complex networks affect individual decision-making, and are there local strategies that will both increase individual player profits as well as stabilize the global supply throughput?
We will extend existing models on the two-player supply chain game to multi-player supply chains with non-trivial connectivity structures modeled via graph theory, and investigate various player dynamics (e.g. consensus, best response, gradient descent) in combination with different interconnection structures(e.g. trees, small-world network, star) to study the stability of the overall supply chain.
As modern technology digitizes sectors including agriculture, power grids, and transportation networks, a network of interconnected and self-interested decision-makers emerges as a prominent feature of urban supply chains. Whereas traditional supply chains tend to follow a multi-tier, tree-like structure, e.g. producers supply distributors, who supply retailers, who supply consumers. Increased global mobility and the formation of retail giants have introduced greater complexity in supply chain pathways. For example, a local bakery may supply a grocery store chain with baked goods, while the grocery store chain may supply the bakery with flour and other baking necessities. How do these complex networks affect individual decision-making, and are there local strategies that will both increase individual player profits as well as stabilize the global supply throughput?
We will extend existing models on the two-player supply chain game to multi-player supply chains with non-trivial connectivity structures modeled via graph theory, and investigate various player dynamics (e.g. consensus, best response, gradient descent) in combination with different interconnection structures(e.g. trees, small-world network, star) to study the stability of the overall supply chain.
This project has the following goals.
1. Learn about supply chain modeling from operation research, game theory under shared constraints, and different graph-theoretical metrics for characterizing connectivity.
2. Formulate a multi-player supply chain game and derive conditions for the existence of a Nash equilibrium.
3. Characterize the stability of gradient descent, consensus and/or best response dynamics against disturbances and investigate their convergence properties.
4. Verify convergence guarantees on a real data set for agriculture supply chains and electricity markets.
This project has the following goals.
1. Learn about supply chain modeling from operation research, game theory under shared constraints, and different graph-theoretical metrics for characterizing connectivity. 2. Formulate a multi-player supply chain game and derive conditions for the existence of a Nash equilibrium. 3. Characterize the stability of gradient descent, consensus and/or best response dynamics against disturbances and investigate their convergence properties. 4. Verify convergence guarantees on a real data set for agriculture supply chains and electricity markets.
Please send your resume/CV (including lists of relevant publications/projects) and transcript of
records in PDF format via email to {huilih,degiulia, shall\}@ethz.ch.
Please send your resume/CV (including lists of relevant publications/projects) and transcript of records in PDF format via email to {huilih,degiulia, shall\}@ethz.ch.