Combining Simulation & Optimization techniques in a Planning & Scheduling environment
Advantages of simulation
One of the main advantages of simulation is handling uncertainties. Traverse Board can simulate plans multiple times, using statistics to mimic real-life uncertainties. In this way it is possible to find out how robust a plan/schedule is and what the risks are. It is possible to compare plans for instance on its Value at Risk, the Timeliness of a Plan and Resource Utilization.
This will allow to optimize for the current schedule, taking into account future plans and uncertainties. Schedulers are able to evaluate the impact of different schedules/plans, either generated by the optimization engine or by manually altered plans. traverseboard Advanced Planning & Scheduling will quantify the risk a plan is exposed to, for the short, mid and long term.
Create, evaluate and publish plans
Interactive Gantt Charts are used to display plans and activate or deactivate recourses. Planned arrivals and stock profiles can be compared to actual data.
Mimic future data
Traverse Board is able to schedule with ‘known’ short term uncertainties and combine it with longer term uncertainties. For instance a breakdown or bad weather can be known when creating a schedule for the next day. For the mid and long term actual data might not be known. Here the simulation engine will provide an order generator to mimic this unknown future data.
Compare different plans using real life uncertainty
Different plans are evaluated through simulation experiments, enabling a planner to create the optimal solution.
Simulation results will quantify Value at Risk and Timeliness
Value at Risk and Timeliness are two important features for determining the best plans. For supply chain solutions SCOR card reporting can be generated as well, including the robustness of these KPI’s.
Simulation-assisted planning & scheduling projects by Talumis
For LNG (liquefied natural gas) supply chain planning and scheduling a solution has been developed that can create and optimize annual delivery plans. The optimization takes into account future uncertainties like weather events and unscheduled production outages. The resulting plan is both robust (can recover from unforeseen events) and optimal (uses the best delivery strategy based on revenue).
Another example is the on-line monitoring and control of coal storage and burn plan in a power plant. The simulation model calculates, based on on-line (real-time) PLC information, where particular coal qualities are located in the storage facility. Then a simple algorithm is used to create the best short-term burning mix and plan. This plan can than also be simulated to validate the calculated plan.
For the planning of supply vessels Talumis also created a simulation support planning system. The problem of vessel supply can be characterized as a so called “multiple vehicle routing problem with time windows”. There are mathematical optimizers that can calculate the ‘optimal’ solution. The downside is that these optimizers are very calculation intensive, cannot cope with all constraints and can only be used deterministic. Talumis created a combination of heuristics and optimization together with a very detailed simulation model. The simulation, together with the bespoke optimizer, is able to create detailed delivery plans that can be used both strategically (how many vessels do I need) as well as in operations (what is the plan for tomorrow).