Discrete event simulation

We use Discrete Event Simulation (DES) to help inform decision making and investigate system dynamics.

Discrete event simulation (DES) is a type of simulation that considers a system as a discrete collection of events. Each event has a defined effect on the rest of the system. The individual processes that make up a system can be defined in terms of their impact on the system, resource requirements, and trigger (i.e. they may be scheduled, occur at random, or appear in response to another system event). After defining the constituent parts, they can be combined within the simulation to recreate the system from the ground up.

One advantage of this method is its speed and configuration. Unlike continuous simulation models where each slice of time is considered equally, DES achieves results much more quickly by only considering those events that alter the system’s state. Additionally, since the system is defined from the bottom up, changes to ‘low-level processes’ can be tested almost instantly without reconfiguring high-level logic.

Applications

  • Options analysis:
    Test various infrastructure, personnel or resource changes without the need for costly physical trials.
  • Throughput analysis:
    Generate utilisation metrics to identify the bottlenecks within the system.
  • Planning:
    Use a model of a proposed system to develop an understanding of the required resources and timeline.
  • Optimisation:
    Optimise system performance against selected key performance indicators

In practice, simulation can provide support across the entire lifecycle of an asset:

Our approach

We take a consultative approach to DES, and we begin by holding discussions with stakeholders to establish the relevant requirements and behaviour of the system. We then collect critical input data and select a preliminary model, followed by an iterative process in which stakeholders check and validate the logic and behaviour of the model. This approach gives our customers confidence in the model being constructed, not just because they can compare the model’s results to real-world data but because they can also see how individual activities affect the bigger picture.

Once the model’s validity has been agreed upon, it can trial proposals, test assumptions, and investigate the relationship between different inputs and KPIs.