Composition of Optimal Service Workflows with Quality-of-Service Enabled Multi-Criteria Uniform Cost Search Algorithm
This paper proposes a search-based method to partly automate the workflow composition, including the planning and execution stages, with web services in a service-oriented architecture. The proposed methodology models the workflow composition problem as a directed and weighted graph, henceforth called the service-oriented architecture graph, where vertices are associated with the degree of completion of the overall task at hand and edges represent service executions. Edge weights are formulated based on the quality of services as defined by the user. A uniform cost search algorithm is adapted and applied to identify the optimal path based on user input, which constitutes an ordered sequence of service executions, from a given initial state to a user-defined goal state. The proposed approach for service oriented workflow composition was applied to an information-technology domain problem to demonstrate its utility through a simulation study. Simulation results indicated that the proposed methodology is feasible and optimal solutions can be computed within reasonable computational cost bounds.