Efficient algorithms for selection of recovery points in tree task models

Efficient solutions to the problem of optimally selecting recovery points are developed. The solutions are intended for models of computation in which task precedence has a tree structure and a task may fail due to the presence of faults. An algorithm to minimize the expected computation time of the task system under a uniprocessor environment has been developed for the binary tree model. The algorithm has time complexity of O(N/sub 2/), where N is the number of tasks, while previously reported procedures have exponential time requirements. The results are generalized for an arbitrary tree model.<
>