Workflow Management System for DMS

Abstract

Dramatic enrichment of high speed communication technology at affordable costs enabled the resolution of problems related to the limitations of power supply. Therefore, introduction of Smart Grid for managing highly controllable power grids has become indispensable. Its very important part is Distribution Management System (DMS). These sophisticated systems execute a large number of workflows with very high resource requirements. In this paper dynamic, centralized scheduling strategies for allocating DMS workflows are presented and compared. Also, a distributed scheduling algorithm for DMS calculation engine is developed. It is argued that minor investment in the resources and introduction of a hybrid scheduling algorithm leads to the significant boost of the system performance. The hybrid scheduling algorithm is the result of combining centralized and distributed scheduling strategies. Experimental study shows that considerable improvement of overall system performance is achieved by using developed algorithms for designing effective schedulers when make-span and workload are optimized.

DMS Workflows Scheduling Concepts Comparison

“…To meet these expectations, DMS is tightly integrated with other systems like Supervisory Control and Data Acquisition, Customer Information System, Meter Data Management, Weather Information System, Asset Management and Geographical Information System. Such DMS, integrated into the Smart Grid concept [3,29], uses a Grid [2,15,20] to store and process aforementioned large amounts of data so that required functionalities are provided [25]. Each functionality demands execution of different (most commonly dependent) tasks on Grid resources.…”

“…The interaction can happen in any random moment, so the full set of workflows that has to be executed (workflow application) is not known in advance, it evolves in time instead. Furthermore, workflows may be mutually dependent [25] and prioritized [21] (e.g. processing of an alarm event is more important than execution of "what-if" analyses [25]).…”

“…Furthermore, workflows may be mutually dependent [25] and prioritized [21] (e.g. processing of an alarm event is more important than execution of "what-if" analyses [25]). Consequently, the additional complexities are included in the workflow management.…”

“…Furthermore, dependencies between workflows may occur, e.g. execution of DMS functions over a part of a DN is directly dependent on the initial import of that DN part [25].…”

“…This section provides a description of the workflow application model [25] which is further used to define optimization problem.…”