No abstract

We develop an algorithm to solve the Bottleneck Assignment Problem (BAP) that is amenable to having computation distributed over a network of agents. This consists of exploring how each component of the algorithm can be distributed, with a focus on one component in particular, i.e., the function to search for an augmenting path. An augmenting path is a common tool used in most BAP algorithms and poses a particular challenge for this distributed approach. Given this significance, we compare two different methods to search for an augmenting path in a bipartite graph. We also exploit properties of the augmenting paths to formalise conditions for which the solution from subsets of the sets of agents and tasks can be used to solve the BAP with the full sets of agents and tasks. In the end, we evaluate and compare the derived approaches with a numerical analysis. The research is funded by Defence Science and Technology Group through research agreements MyIP: 7558 and MyIP: 7562.

An assignment problem arises when there exists a set of tasks that must be allocated to a set of agents. The bottleneck assignment problem (BAP) has the objective of minimising the most costly allocation of a task to an agent. Under certain conditions the structure of the BAP can be exploited such that subgroups of tasks are assigned separately with lower complexity and then merged to form a combined assignment. In particular, we discuss merging the assignments from two separate BAPs and use the solution of the subproblems to bound the solution of the combined problem. We also provide conditions for cases where the solution of the subproblems produces an exact solution to the BAP over the combined problem. We then introduce a particular algorithm for solving the BAP that takes advantage of this insight. The methods are demonstrated in a numerical case study.

The Lexicographical Bottleneck Assignment Problem (LexBAP) typically requires centralised computation with ( 4 ) complexity. We consider the Sequential Bottleneck Assignment Problem (Se-qBAP) and its relationship to the LexBAP. By exploiting structure of the Bottleneck Assignment Problem, we derive an algorithm that solves the SeqBAP with ( 3 ) complexity. We analyse conditions for which solutions sets of the LexBAP and the SeqBAP coincide. When these conditions hold, the algorithm solves the LexBAP with computation that can be distributed over a network of agents.

No abstract