Scheduling and cost optimization of investment, development and maintenance tasks and optimal resource assignment for E.ON.
Large-scale utility networks, such as a gas transportation pipeline or electricity transmission lines, require continuous maintenance and development to function adequately. Some parts of the transmission network shall be monitored regularly, while other components require scheduled maintenance when necessary. Based on the maintenance and development policy of the network operators, a set of future tasks could be obtained from the ERP. Each task has a priority and clearance level, a spatial location, a temporal constraint determining when the maintenance can be performed. Various additional side constraints may also limit possible scheduling or assignment. In order to facilitate the implementation of future tasks, both an internal capacity of resources and a pool of external contractors are available for scheduling and assignment.
These resources have a spatial location, temporal availability, a pre-existing schedule already known at the time of planning, a clearance level, a capacity and a cost function, which determines the implied cost of assigning this certain resource to an arbitrary task. In case of multi-objective optimization and planning, our primary goal is to minimize the total cost of future tasks by scheduling the resources and assigning them to one or more tasks, while obeying business constraints. Our secondary goal was to provide an analytic tool in which certain aspects of a scheduled plan could be fixed in order to assess and evaluate different ’what if’ scenarios.
The maintenance scheduling and resource assignment plans made by our solution should adhere to the strategic goals of the network operator as well. If these plans are of insufficient quality then either the total cost will exceed the cost limit due to the underutilization of the available resources, or the number of scheduled tasks will not reach the threshold in a given timeframe, which leads to suboptimal maintenance and a possible failure of the underlying network.
NEXOGEN approached E.ON with a custom-tailored decision support solution based on a proprietary mathematical optimization framework. This framework was able to effectively address the issue of delivering high-quality maintenance task schedules and resource assignments. The goal was to introduce an integrated solution to the existing IT infrastructure to make optimal plans, satisfying business constraints in a short timeframe. A secondary objective was to save error-prone and time consuming activities for the decision makers, which could otherwise hinder the efficiency of the underlying decision making pipeline.
Analyzing the domain problem was our first step.Our proficient business analysts identified and assessed the business processes arising in the context of gas transportation and electricity transmission network maintenance planning at E.ON. In the very early stage of the project, critical factors and key performance indicators of the underlying problem were determined. The scheduling of tasks and the minimization of the total cost of task and resource assignments were necessary to make the overall project successful and efficient from both technical and business standpoints.
After this stage complex mathematical models and accompanying algorithms have been developed to efficiently describe and solve the NP-hard combinatorial optimization problem of simultaneous resource and task scheduling and the generalized resource and task assignment. This modeling process also involved the incorporation of various side-constraints and objectives imposed by the domain problem. These models and methods have been embedded into the optimization framework of NEXOGEN. The solution was integrated into the corporate IT infrastructure of E.ON, which is based on an SAP ERP solution, via standardized interfaces. The analytic tool was implemented in an ASP.NET MVC environment. Regardless our system has clear and well-documented interfaces, which can be connected to any major ERP platform/vendor, such as SAP, Microsoft Dynamics AX and Oracle.
At the end the integrated solution has evolved into a high performance, custom-tailored decision support system and optimization tool, which suits the needs of E.ON the best and is able to obtain high quality, executable, valid and optimal simultaneous task and resource scheduling and assignment plans. The integration of this solution into the customer’s planning pipeline resulted in substantial improvement of the key performance indicators, which could be measured in explicit cost savings. The solution paid off within half year after its introduction, while the entire project, from the start of development to the go-live phase, took no more than 7 months in total.
© 2018 / created by: pixelephant