Discrete-event simulation has been around for a long time. Growing computation power has aided the application of discrete-event simulation in recent years. This is especially true with regards to visualization and animation of the models themselves. However, the method, related procedures, the process and software related to discrete-event simulation was also applied 20 years ago. Since then, many simulation projects have helped business improve their layouts, processes and grow their profits and customer satisfaction. But, a lot of projects failed. In today’s blog post I comment on commercial simulation project failures.

Failure rates of discrete-event simulation projects

It is difficult to make a generalization about the failure rate of discrete-event simulation projects. Nevertheless, various studies have been released on this subject. Their estimate of project success rates range from 40% to 75%.It is generally believed that roughly half of all commercial simulation projects in the supply chain management and production planning domain fail to meet basic expectations of relevant stakeholders. This success rate is generally believed to be significantly below the average success success rate of analytics projects in supply chain management and production planning.

Traditional analytics projects generally involve the use of statistical and data analysis techniques to gain insights and make data-driven decisions. Operations research projects, on the other hand, furthermore make use of mathematical models and concepts mainly developed for operations improvement. This e.g. comprises mixed-integer programming and optimization, constraint programming, non-linear optimization, heuristic, genetic optimization, and discrete-event simulation.

While traditional analytics and operations research projects can be complex, they often involve relatively static data sets and can be completed using analytical tools and techniques. As a result, the success rate of analytics projects can be relatively high, particularly when compared to DES projects. DES projects, on the other hand, require a deep understanding of complex systems and processes, as well as the ability to develop accurate and effective simulation models. These projects involve dynamic data and require more advanced modeling and simulation techniques. This high level of complexity may be one of the major reasons for why simulation projects have a higher failure rate than traditional analytics.

Root causes for simulation project failures

Effective project management practices, including thorough planning, stakeholder engagement, and effective communication, significantly influence the success rate of commercial simulation projects.

Simulation projects can fail for a variety of reasons, but some of the most common root causes of simulation project failures include:

1. Lack of clear project goals
2. Poor data quality
3. Lack of expertise
4. Inadequate testing, verification, and validation
5. Insufficient stakeholder involvement
6. Poor communication
7. Scope creep
8. Unrealistic assumptions
9. Unrealistic expectations
10. Lack of resources

By addressing these root causes, simulation project teams can improve their chances of success. Simulation teams and engineers must thus continuously improve their project management and communication skills, technical proficiency, and stakeholder management.

Important skills to avoid simulation project failure

Discrete-event simulation (DES) departments in a supply chain management or production planning setting require a range of technical and non-technical skills to be successful. Some important skills include:

• Technical proficiency
• Data analysis
• Process mapping
• Communication
• Project management
• Problem-solving
• Teamwork

On the technical end of the spectrum, the following skills should be required and improved:

• System knowledge, i.e. understanding of the production and logistics processes as well as therewith associated relevant equipment and systems
• Statistical analysis, i.e. to be able to evaluate the confidence of results deducted from the simulation and the ability to derive generalizable findings from the simulation results
• Data analytics – obtaining, wrangling and visualizing data
• Systems thinking, i.e. ability to conceptually model a complex system
• Object-oriented programming
• Software engineering skills

Related content

You can learn more about discrete-event simulation and its applications on SCDA:

• Link: Visual Components financial KPI simulation
• Link: Open-cast mine simulation for better planning
• Link: Manufacturing simulation for plant design
• Link: Discrete-event simulation (DES) use cases
• Link: Discrete-event simulation software list
• Link: Manufacturing simulation

Linnart Felkl

Data scientist focusing on simulation, optimization and modeling in R, SQL, VBA and Python