A simulated annealing algorithm for dynamic document retrieval

A simulated annealing procedure is proposed as an alternative for scheduling document retrieval from retail bank offices. The new scheduling tool yields high solution quality more efficiently by addressing performance and implementation problems with random search methods; trapping at local minima and path dependency. Performance comparisons are based on sample problems generated to resemble those in the application environment. Significance: New information technology provides near real time access to retail bank transactions data that can be exploited in scheduling document retrieval couriers. Simple but efficient algorithms can apply this information to significantly reduce delays in document retrieval. This paper describes how these algorithms have been applied to improve on manual and random scheduling procedures.