Lattice-Valued Bottleneck Duality

Robert Ghrist,Julian Gould,Miguel Lopez
2024-10-01
Abstract:This note reformulates certain classical combinatorial duality theorems in the context of order lattices. For source-target networks, we generalize bottleneck path-cut and flow-cut duality results to edges with capacities in a distributive lattice. For posets, we generalize a bottleneck version of Dilworth's theorem, again weighted in a distributive lattice. These results are applicable to a wide array of non-numerical network flow problems, as shown. All results, proofs, and applications were created in collaboration with AI language models. An appendix documents their role and impact.
Optimization and Control,Combinatorics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to reformulate the classical combinatorial duality theorems in the framework of order lattices, especially for the bottleneck path - cut and flow - cut duality problems in networks with distributive - lattice - valued capacities. Specifically: 1. **Extension of classical flow - cut duality**: The paper aims to generalize the max - flow min - cut (MFMC) theorem to the case where edge capacities take values in a distributive lattice. The classical max - flow min - cut theorem states that in a network, the maximum flow from a source to a sink is equal to the capacity of the minimum cut. The paper generalizes it to values on a distributive lattice. 2. **Extension of bottleneck duality**: In addition to the classical flow - cut duality, the paper also focuses on bottleneck duality. Bottleneck duality considers the relationship between the value of the minimum - capacity edge on a path and the value of the maximum - capacity edge in a cut - set. The paper generalizes this result to the case of distributive - lattice - valued capacities. 3. **Bottleneck version of Dilworth's theorem**: The paper also generalizes Dilworth's theorem to the case of distributive - lattice - weighted. Dilworth's theorem describes the duality relationship between chains and anti - chains in a poset. The paper proposes a bottleneck version of Dilworth's theorem, which considers the relationship between the maximum of the minimum - weight elements in a chain and the minimum of the maximum - weight elements in an anti - chain. ### Formula representation To understand these problems more clearly, the following are some key formulas involved in the paper: - **Max - flow min - cut theorem (MFMC)**: \[ \max_{\phi \in \Phi} \text{val}(\phi)=\min_{C \in C} \text{cap}(C) \] where: - $\Phi$ is the set of all feasible flows from the source $s$ to the sink $t$. - $C$ is the set of all cuts that separate the source $s$ and the sink $t$. - $\text{val}(\phi)=\sum_{v:(s, v)\in E} \phi(s, v)$ represents the value of the flow $\phi$. - $\text{cap}(C)=\sum_{e\in C} c(e)$ represents the capacity of the cut $C$. - **Bottleneck duality theorem**: \[ \max_{P\in P} \min_{e\in P} c(e)=\min_{C\in C} \max_{e\in C} c(e) \] where: - $P$ is the set of all paths from the source $s$ to the sink $t$. - $C$ is the set of all cuts that separate the source $s$ and the sink $t$. - **Bottleneck duality theorem on distributive lattices**: \[ \bigvee_{P\in P} \bigwedge_{e\in P} c(e)=\bigwedge_{C\in C} \bigvee_{e\in C} c(e) \] where: - $c(e)$ is the capacity of the edge $e$, taking values in the distributive lattice $L$. - $\bigvee$ and $\bigwedge$ represent the join and meet operations in the lattice respectively. Through these generalizations, the paper provides a new perspective for dealing with non - numerical network flow problems and shows the potential applications of these results in multiple fields.