Scheduling is used for daily order termination in the production environment, ensuring near-optimal resource allocation and adherence to deadlines for most customer orders.
What scheduling systems are for?
Differences in scheduling systems
Impact of variability, an experiment
We offer simulation or optimization based scheduling technology. Both of them can:
schedule production orders forwards
experiment with additional customer demand
take into account limited capacities of resources
plan material usage with bill of materials
consider and calculate priorities, qualities or other factor
depending on resource availability
manage qualifications of the personnel
use changeover matrices for machines
take into account alternative resources and operators
apply shift models
calculate secondary resources (such as tools with limited qty)
visualize results in Gantt charts
report with detailed log lists per resource
bind directly to databases or other data sources
In other points differ the two scheduling techniques considerably. It is important to realize their advantages and disadvantages in order to use their potential correctly.
Evaluates the production environment dynamically with detailed transport activities, delayed parts supply and congestion in production.
Bottlenecks can be specifically managed and planned, so that all other processes follow them accordingly.
Calculates and shows the likelihood of order completion, takes into account random delays that can happen in production such as equipment failures or unavailable staff.
Precise availabilities and movements for staff, batches and resources.
Builds up all unique features in a production system faithfully.
3D animation allows for easy validation of the daily plan.
Each project is worked out in detail using simulation models, with many aha effects for all involvedunique advantages.
Reverse planning strategies possible.
Orders can be manually fixed at resources.
Seeks global best combination of processes in production according to a user defined target function. Thus, the orders are resorted and individually scheduled, until the target function achieves the best result.
There are predefined data masks and solid model framework for many standard applications in the production environment.
Includes a user console with active feedback loop from production. In case of discrepancies it can be rescheduled on the fly.
Visualizes order networks and order dependencies.
Large costumer pool.
An experiment, which scheduling type can predict the delivery dates more precisely?
We have compared the two types of scheduling with each other in a simulation experiment. For that we have built a simulation model with a dynamic production process. The process steps times are stochastic times in a certain range (min, max) rather than fixed times. The resources in the model may also fail randomly.
The simulation-based scheduling utilizes directly the process time, resource variations and failures in the model.
The quasi-optimization-based scheduling, however, requires fixed process times so we provide an average process times instead and remove resource failures from the model.
In this experiment we let both scheduling tools create their schedules for all the production orders first. Then we run the simulation model again to find out a possible quasi production period outcome. At the end of the simulation run we have got all the results painted in one chart. The chart shows the sum of delays in hours for the orders completed by the time of every measurement.
Orange line – due dates – shows the target values, the delivery dates promised to customers (same as due dates, so no delays).
Brown line – deterministic – the optimization based scheduling shows a value of about -1400 hours (minus 1400) at the end. The sum of all orders delays is negative. On average are the orders produced much earlier than their due dates.
Blue line – stochastic – the simulation-based scheduling ends at about + 1700 hours. The orders were very delayed. On average they are much later than their due dates.
Green line – the actual seguence in production-ends at about + 1000 hours. Orders were on average late.
The simulation-based scheduling can be more effective in a dynamic production environment than the optimization-based scheduling. It predicts delivery dates more accurately, because many potential process variations and possible production events are considered very early in the planning.
On the other hand the production orders are processed sequentially in the simulation-based planning according to their start date and perhaps some local selection rules. Contrary to that the optimization-based planning utilizes a global search strategy for the best sequence of the individual orders and operations. That is why the optimization-based planning is usually preferred for well predictable production systems, i.e. with little variability.
We can show you on your system what each of these scheduling types means for your production system. Please contact us.
Tell us what your project is about and we help you find a good solution.