Resource-to-task allocation is often done in near real-time due to the latency of the task information, sudden task arrivals and other uncertainties. This means that the problem must be tackled within a very tight time budget.

Local search-based heuristics are particularly adequate for this problem due to their low computational complexity and ability to generate good quality solutions.

To address these issues, we modelled the problem as a series of assignment problems and designed multiple local search algorithms to tackle these problems sequentially. These algorithms and examined and analysed empirically to evaluate their suitability based on the problem features.