Core Concepts
The worst-case degradation of social welfare when allocating connected subsets of items is quantified through the concepts of egalitarian and utilitarian price of connectivity.
Abstract
The paper studies the allocation of indivisible goods that form an undirected graph and investigates the worst-case welfare loss when requiring that each agent must receive a connected subgraph. The focus is on both egalitarian and utilitarian welfare.
Key highlights:
The authors introduce the concepts of egalitarian and utilitarian price of connectivity, which capture the worst-case ratio between the optimal welfare and the optimal connected welfare.
For the two-agent case, they provide tight or asymptotically tight bounds on the price of connectivity for various graph classes, including complete graphs with a matching removed, complete bipartite graphs, and graphs with connectivity 1 or 2.
For the three-agent case, they show that the egalitarian price of connectivity for tree graphs is equal to the maximum number of disjoint connected subgraphs when two vertices are removed.
They also extend their results to the general case with any number of agents for certain graph classes like stars, paths, and cycles.
The results are supplemented with algorithms that find connected allocations with a welfare guarantee corresponding to the price of connectivity.