The electrification of society and the rise of renewable energy require an enormous expansion of the electricity grid. A major challenge in achieving this is the shortage of technical personnel. As a result, a large Distribution System Operator in the Netherlands is unable to carry out all the projects in its high-voltage network that it would ideally pursue. The organisation therefore needed to develop a realistic plan for all projects over the next ten years.
A complex planning puzzle
The complexity of this planning lies in the nature of the projects themselves: each can span several years and consists of multiple phases, each requiring different skill levels. Furthermore, some resources are available only in specific regions, while others can operate across the entire service area. Projects vary in priority, start date, and deadline, and at the same time may also be related to each other. For instance, one project may only be completed once another has finished.
The existing approach was to create the planning by hand. With hundreds of projects to consider, this process took a lot of effort and time. In this manual approach, certain constraints such as interdependencies between projects were ignored to reduce complexity, and simple rules were applied to determine the allocation of projects to years, resulting in suboptimal outcomes.
Unlocking the potential of a model
In multiple iterations, we developed a model to automate and optimise the planning process. The model respects all constraints, and the resulted solution is an optimal one. In the planning process, there are multiple, possibly conflicting objectives. For example, maximising the number of high-priority projects completed versus meeting as many deadlines as possible. The user can specify the importance of each objective by ordering them.
The model does not only deliver better solutions, it is also far faster than the manual method. What previously took hours is now resolved in minutes. This makes it practical to run the model multiple times and test different scenarios, for instance to evaluate the consequences of a different way to organize the mechanics.
Solving general planning problems
While this case is tailored to the specific context of this DSO, many of its elements such as deadlines, priorities, and resource constraints are universal. Project planning is a problem at which mathematical models excel, and one where they can serve as a powerful decision-support tool. If you are wondering whether that might also be true for your planning challenge, feel free to reach out at info@doingthemath.nl.