Manufacturing Informatics

Greybox Schedule Optimisation

Greybox offers a scheduling solution to manufacturers seeking maximum value from their resources. It is a joint cognitive system where human and machine intelligence collaborate. The results are exceptional.

 gb

Up to 40% Labour Reduction

f

10% Output Improvement

Estimated 15% Energy Savings

Up to 40% Tool Set up/Product Changeover Reductions

Current scheduling technologies are inadequate

The two technologies currently used for scheduling both have advantages and limitations.

Humans and/or excel-based templates Black box optimisation algorithms
Has done a “good” job Does a “very good” job
Often not good enough, especially with the onset of mass customisation/one-off products Can comprehend all of system that can be entered into data model
Flexible Inflexible – minimal interaction allowed
Limited in scope and speed of response Requires extensive training for planners to use and sustain
Does not clearly represent business KPIs
Is difficult for end users to understand and amend
Relies completely on the data in underlying systems
Does not comprehend nuances

Greybox Scheduling – beyond state of the art

Creates a joint cognitive system for businesses where human and machine intelligence collaborate for success.

The system's algorithms can suggest solutions to support the end user, increasing user confidence and enabling easier deployment

It offers a clear and simple user interface with a focus on the manufacturing KPIs

User interaction is a key component, allowing access to human expertise and tacit knowledge

It is robust at handling system changes

Intuitive Complex Schedule Management

Merging Human Expert Knowledge and Computer Computational Power

How Greybox works

Observability & interaction is necessary to increase user’s confidence in decision support systems.
This requires seamless transition between human and automated planning.

Optimisation Visualisation
Multi-objective, genetic algorithm optimisation of production and energy Optimisation outputs are visualised as a weekly run-plan
Ability to accept ad-hoc user defined goals based on expert knowledge Users can intuitively generate dynamic rules by dragging and dropping events
Scheduled updates and re-optimisation triggered by critical events Interaction with the run-plan automatically updates projected performance indicators
Scheduled updates and re-optimisation triggered by critical events