Optimisation

When you sign up for the TilliT AI Scheduler, you have a range of options for Optimisation. Please contact our account executives to discuss Optimisation options suitable for your business.

If you have already been licensed for a particular Optimisation algorithm, then please see below for configuration options for each algorithm in the TilliT Optimisation Library.

Create an Optimisation Profile

Optimisation processes can be configured using Optimisation Profiles.

An optimisation profiles defines how long an optimisation process is allowed to run, the size of the population (for evolutionary algorithms) and how important each objective is, among other things.

TilliT optimiser comes with 3 default profiles: Balanced, Detailed and Efficient.

Click on the + button to create a new Optimisation Profile

The Changeover, Makespan and Overdue are seen as competing objectives that you are trying to balance when optimising. Adding weighting to them will configure how much more a feature of scheduling is more prioritised to you.

The weighting works in relative to each other. For example, changeover_weight=2.0, makespan_weight=1.0 means changeovers are twice as important.

Field	Description
Changeover weight	Purpose: Controls the priority/penalty for equipment changeovers in the optimisation algorithm. Higher weight (>1.0): Algorithm tries harder to minimize changeovers by grouping similar products/operations together Lower weight (<1.0): Algorithm is more flexible about changeovers in favour of other objectives like due dates Use case: Manufacturing environments where changeovers are costly, time-consuming, or create waste
Makespan weight	Purpose: Controls the priority for minimising the total completion time (makespan) of all scheduled jobs. Higher weight (>1.0): Algorithm prioritises finishing all work as quickly as possible Lower weight (<1.0): Algorithm may accept longer total completion times to optimise other objectives Use case: When meeting overall project deadlines or maximising throughput is critical
Overdue weight	Purpose: Controls the penalty for jobs/orders that become overdue relative to their due dates. Higher weight (>1.0): Algorithm strongly avoids late deliveries, even if it means less efficient scheduling Lower weight (<1.0): Algorithm may allow some late deliveries to achieve better overall efficiency Use case: When customer satisfaction and on-time delivery are paramount

Using the Optimisation Profiles

Once the Optimiser has been configured it can be enabled in the Scheduling screen by clicking the Optimise Button in the Scheduling View:

You will be presented with the following dialogue:

Go to the Using the Scheduler section for more information on using the Optimiser to automatically create optimised schedules.