By Isabelle Tardif, M.Sc., CQE, CSSBB, ASQ Senior Member. On February 26th, we had the opportunity to familiarize ourselves with How Simulation Can Enhance Scheduling by Avoiding Lean waste.
Mr. Vincent Béchard, the presenter, has been a decision-making consultant in the areas of modelling, simulation and optimization of industrial operations since 2004. He knows simulations very well and shared this powerful tool with us. Understanding the universe of simulation can at first glance sound scary, but Mr. Béchard took us step by step, and built our confidence on the subject.
He started by asking about our experiences with scheduling and how often had we seen a schedule unfold exactly as planned? With some well-chosen examples, we realized that waiting time, queuing, unnecessary motion, inventories piling up, are often part of the real life and are not planned for. Also, although schedules are generally based on average time, their duration will vary around that average -- sometimes you will beat the schedules and sometimes you won’t. Therefore, a schedule that plans with a 100% resource utilization will suffer when variability happens.
The scheduling process itself hides some waste too: often errors in formulas happen in spreadsheets revised for the 100th time, or errors in calculations or loss of creativity from planners too busy to think of better ways to work because it takes so much time to generate a schedule.
Once the awareness on waste and scheduling process was established, simulation arrived to the rescue!
Simulation is hard bolted with Lean, statistics and programming. We were introduced to Discrete Events Simulation, events based modelling models systems that change states (example for a machine state: idle, setup, processing, down) at discrete points in time as a result of a specific event.
To simulate, you first need to build a model that reflects real life. In addition to a simulation software, for discrete modelling, inputs to the model are:
- Design specifications: layouts, routes, dimensions, equipment characteristics, technical data, etc.;
- Operations knowledge: cycle times, schedules, resources availability, operational constraint, etc.;
- Historical process data: not averages of the data, but its distribution.
Once the model is built, it needs to be validated by experimentation and with feedback from experienced people in the organization. With their input, the model will be optimized to the point that it will reflect more and more accurately the process. Output of the simulation includes performance diagnosis, effective and ultimate throughput, equipment utilization rates, meaningful and custom statistics, trends, pie charts, and histograms in a feasible plan. It is a powerful tool to quickly all the scenarios you want to try and so choose the best combination.
To demonstrate the power of simulation and its benefits, Vincent presented a case of an underground mine construction in Canada. Initially, three schedules were made by the mine teams:
- Life-of-mine plan;
- Development plan;
- Procurement and logistics plan
Even with those plans, there were still a lot of mysteries and many aspects to consider that each plan individually did not address. Furthermore, there were signs that the project was late and that the five-year project would not be respected. After two days in the war room with experts representing three disciplines, the decision making project manager and with M. Béchard leading the simulation, they modified several lay-down capacities and locations, modified the drift development schedule and swapped logistics sequences. Results were impressive: they caught up the initial delay, reviewed the final plan to complete construction in five years and as a bonus, found a path to complete it nine months early.
This was a huge project. Smaller projects in other industries (even in health care) could use simulation in their schedules to get benefits and reduce Lean waste. Finally, as Mr. Béchard said: “Even the best simulation will never be able to perfectly and every time encompass subtleties and complexity of real life, but it could for sure save you a lot”.
Next time you think about how to improve scheduling, why not look for simulation in your tool box?