A Deterministic Approach to Nurse Rerostering Problem

When a nurse at a hospital is unable to be in charge of any of his/her assigned shifts, the roster should be rescheduled to cope with it, which is called nurse rerostering problem. In this problem, we are supposed to find a new one which is close to the given as well as satisfies all the constraints that it inherits from nurse scheduling problem which built the roster. It was proved NP-complete so that many heuristic algorithms were proposed. By treating the difference between two rosters as a constraint of the highest priority, it can be thought of as a shortest path problem. In this study, we applied a deterministic algorithm, iterative deepening depth-first search, which was proved to be asymptotically optimal along all three dimensions for exponential tree searches: the amount of time it takes, the amount of space it uses, and the cost of a solution path. The results of the experiments show that it is of practical use for the problem. In terms of the number of differences, it provides us with the optimal solution for all the datasets. Based on the results, the algorithm can be applied to the problem for limited depth or time. If a solution is not found within the limit, we can apply other heuristic algorithm such as simulated annealing or genetic algorithm to find an approximate.