This program takes up a large amount of memory, I have used up to 150MB of virtual memory before finding the optimal solution. It would be helpful if some of the memory that is holding partial schedules that have been made too costly by the last assignment of lowest to be removed from the heap. This would have the effect of reducing the amount of memory needed and making it possible to search longer tasklists. But when a complete schedule is found that is shorter than the initial upper bound it is the optimal schedule, I haven't seen a case where this hasn't happened although it doesn't have to. Because the only time a cleaning can occur is after a new upper bound is found, and it is very rare that another will be found that is lower it makes little sense to clean the heap. Any cleaning of the heap with this characteristic of the algorithm would just mean emptying the heap to exit the program, which is done by the algorithm already.