This chapter will help you analyze the performance of your process before your process goes live in a production environment. With analysis results, you can change the parameters of your process so that the total cost and time of running a process for a business can be reduced. You will use simulation after implementation, since you should not consider simulation as a process design tool, but instead as a tool for optimizing a processes performance before it goes live. In this chapter, you will cover the following topics:
- Defining simulation models
- Defining simulation definition
- Running a simulation
- Analyzing simulation results
- Re-engineering the BPM Process to improve performance
Simulation is an important function in Business Process Management, as it answers some of the important questions specific to process usage, such as the following:
- How much the process will cost the business
- Predicting Business Process behavior
- Checking if the process output meets the process objectives
- Identifying bottlenecks, if any
- Effect of changes on an existing process design
Simulation can be performed before implementation, but here we will be using simulation after implementation. You can perform simulation anytime before or after implementation, however it should always be before you go live with your process. Readers can perform simulation before implementation with the methods given in this chapter. However, while reading through this section, they will understand why simulation is used for analysis after implementation.
Process Modeling focuses on getting the process right before implementation. Many use simulation before implementation because they think that direct experimentation with the business process would be disruptive or costly.
Simulation as a process design exercise makes more sense when the costs of implementation are relatively high as compared to the costs of simulation analysis. However, at present, the cost of simulation has come down significantly when used for analysis after implementation.
You are not at all delaying the benefits of implementation by positioning simulation for use after implementation. Implementation before simulation is helpful for many reasons, such as the following:
- Implementation actually provides the data required for parameterized simulation models virtually for free
- High quality of data is easy available
- Accurate data is guaranteed
- The potential for process discovery and the skills required to create meaningful simulation models are reduced, as the simulation model construction process is partially automated as a consequence of Process Implementation.
- Direct data usage increases the predictive ability of simulation models
At whatever phase you use simulation, the question is why you need it. This is because, as an Analyst post-modeling, or as a Developer post-implementation, you want to optimize the process before it goes live, so that you can reduce the cost a process brings to the business, and so that Service Level Agreements (SLA) are met. Oracle BPM Simulation addresses structural improvements most common in process redesign. It should provide useful cost and quality metrics, in addition to time-based metrics. It actively assists the user with setting model parameters to match known aggregated metrics of the as-is process.
Simulation does not execute each individual task within a process. That is, the code within an activity is not executed, forms are not displayed to the user, variables are not assigned values, external resources are not updated, and so on. However, you can mimic the behavior of an activity by configuring different attributes within a simulation model. These attributes include duration, resources, costs, queue information, and probability of sequence flows.
Process simulation models specify parameters such as the following:
- The rate at which new instances will be created
- The estimated execution times for each of the activities
- What percentage of the time a gateway will take one branch or another and so on
So, you can use the tool to create different 'what-if' scenarios based on different combinations of resource allocation and activity behavior. You can even run simulations from multiple processes simultaneously, to identify resource contention across multiple processes. Once you run the simulations, you can examine the results too.
The process of defining simulation for a BPM process comprises of defining the simulation model, defining simulation definitions, and running the simulation.
Although simulation can be used right after the modeling phase, you as a Business Analyst can analyze the process and its performance and define the operational cost to business. You will carry out the similar task as a Developer, post implementation. However, you are carrying out this step much before your process goes live in production. You have assumed that modeling, implementation, and deployment done up to this point are all carried into development instances.