Scheduling of tasks using a bayesian network

A computer program product comprising a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code configured to implement a method comprising:

• creating a graph comprising nodes and edges, the nodes being associated with a set of tasks respectively, wherein each edge of the graph indicates an execution relationship between the set of tasks of two nodes related by an edge;
• reversing a direction of each edge of the graph for indicating a reversed dependency, resulting in an updated graph;
• determining for each node in the updated graph a conditional probability table using probabilities defined using a first failure rate of a task of a node and further using a second failure rate of a task of a parent node of the node if the node is a child node;
• performing a Bayesian inference to compute for each node of the updated graph a respective marginal probability using a respective conditional probability table;
• and
• using the respective marginal probability for each node to determine an execution plan having a highest probability to fail sooner.

A computer system for scheduling execution of a set of tasks, the computer system being configured for:

• creating a graph comprising nodes and edges, the nodes being associated with the set of tasks respectively, wherein each edge of the graph indicates an execution relationship between the set of tasks of two nodes related by an edge;
• reversing a direction of each edge of the graph for indicating a reversed dependency, resulting in an updated graph;
• determining for each node in the updated graph a conditional probability table using probabilities defined using a first failure rate of a task of a node and further using a second failure rate of a task of a parent node of the node if the node is a child node;
• performing a Bayesian inference to compute for each node of the updated graph a respective marginal probability using a respective conditional probability table;
• and
• using the respective marginal probability for each node to determine an execution plan having a highest probability to fail sooner.