The framework below is a representation of what a good Next Gen production scheduling process looks like:

1. It starts with the inputting of data from different organisational areas: maintenance plans, real-time machine conditions, demand and supply information. If unexpected changes occur (e.g., customer order changes, machine breakdown), they are immediately reflected in the data model, so that it is possible to adapt the plans promptly. In particular, the triggers considered are external and internal variables (e.g., labour strike) that alert the system to the need to change the schedules.
2. The optimisation model runs different scenarios according to the current situation (inputs, parameters, constraints) and the defined objective functions (e.g., maximise resource usage, minimise changeover time).
3. The schedules are then simulated on the digital twin which is seamlessly connected to the optimisation model, where the whole process is tested, and the machines’ behaviour and health status are analysed. In this way, the robustness of schedules is verified, helping to prevent unplanned shutdowns and defective products as a result of a degraded machine status.
4. Then the information from the simulation is fed back to the optimisation model so that the best schedule can be identified and suggested to production planners.
5. Finally, planners review the results and analyse the KPIs to approve or modify the optimal schedule, considering the impacts on the plant and the overall supply chain.

Real Time Data Availability

Data is the ‘fuel’ of Next Gen production scheduling, and it is complex to tackle. The first requirement is that the relevant data points (e.g., data from which machine), granularities and frequencies (e.g., real-time or periodical) must be defined.

Once this is clear, the next step is to identify all the data stakeholders and connect the relevant systems to the overall ecosystem and database. For instance, the inventory status and workforce availability are information stored in different systems which must be harmonised to allow the ‘live’ data feed.

Based on the requirements of the company and the characteristics of the business, the data strategy must consider and define policies to allow this information from the live feed to change schedules. For instance, a last-minute customer order change could compromise the whole schedule and it might be necessary to reschedule. However, this should only be permitted if a critical economic or operational threshold is met.

Visualisation

The easiest way for business users to take informed actions rapidly is by using visual support. Visualisation is a great asset to improve decision-making performance, in terms of responsiveness and efficacy, and foster adoption of the system.

Production planners can access the digital twin of the plant with different levels of granularity to perform root cause analyses that drill down to the shop floor. In this way they can identify bottlenecks early and swiftly take corrective action. Ease of access to the plant's information ensures swift decision-making.

Advanced Analytics

The goal of the Next Gen production scheduling optimisation model is to offer feasible recommendations based on business objectives (e.g., profitability) in a short timeframe.

The optimisation model can be seen as an engine with different power levels; it needs to be ‘refuelled’ constantly with data and can be built on different optimisation methodologies (heuristics, meta-heuristics, reinforcement learning).

Its main activities are scheduling, rescheduling, and reallocating jobs. The aim is to provide reliable support to planners who often struggle to define the best plan without testing it and knowing its impacts.

A sophisticated optimisation model analyses a large quantity of data and considers financial and sustainability policies (e.g., do not emit more than a certain percentage of CO2 per schedule), demand requirements, and supply constraints (e.g., order fulfilment dates), to define the most cost-effective solution.

Optimisation needs a robust simulation model to test scenarios and predict their impacts based on the real environment, thereby enabling planners to make the best decision.

Dynamic Master Data Maintenance

The interconnection of systems is crucial. It opens up scope for intelligent operation by leveraging dynamic data.

This includes machine specifications and attributes such as age, performance, scrap rate and the maintenance performed so far. It is critical that roles and responsibilities for data quality and maintenance activities are assigned within the organisation to supervise and manage exceptions even if the process is automated.

For instance, when a machine is used for a long time, it starts to degrade and its performance drops dramatically, resulting in fewer good products (i.e., a higher scrap rate and greater CO2 emissions per output). This information has to be automatically stored and updated in the machine master data specifications, resulting in dynamic data maintenance that allows these inputs to be included when scheduling production.

Network visibility

Once the shop floors are fully integrated and digitalised into the digital twin, planners finally have a 360° view of what is happening in the plant. Thanks to the great amount and quality of data collected, the visualisation capability can be leveraged as a reliable means to identify production bottlenecks at a glance.

Interconnected Next Gen production scheduling across plants provides site visibility and the ability to mitigate risks across the supply network. For instance, if a specific plant experiences a labour strike that prevents it achieving the production goals, then accessing another plant’s capacity information and reallocating work there could be a strategic asset for the company, mitigating disruptions.

This harmonised inter-connection that provides visibility between plants in different countries could also lead to benchmarking activities for continuous improvement initiatives (e.g., through comparison of financial and operational performance).

Preparing the foundation involves acquiring, processing and modelling the relevant data to synchronise the digital twin. There are two types of data needed to create your Next Gen characteristics (on top of transactional and master data):

1. Machine data is collected through IoT sensors, processed and distributed more efficiently through the cloud and edge computing. It is critical to obtain inputs from subject matter experts to:
   1. analyse the most critical machinery
   2. understand which mechanism/movement to analyse
   3. interpret machine data (e.g., mechanical, and electrical machines have different degradation processes)
2. The data related to the plant layout is collected by scanning and mapping the site, and then through 3D modelling the plant is digitalised. By allocating machine data in the plant representation, the digital twin is active and synchronised real-time with the physical plant. With this, the business user can visualise the entire plant production to a high level of granularity and follow the alerts by drilling down to production lines to detect anomalies and bottlenecks. This means time is not wasted investigating potential root causes.

When it comes to dealing with disruptions to the production schedule, optimisation is needed to identify robust scenarios that meet the business objectives and mitigate unexpected constraints. It is fundamental to identify the most suitable model and settings based on the company’s characteristics and business needs (e.g., financial targets, sustainability policies) and calibrate its level of sophistication, i.e., the time required to provide solutions, based on the responsiveness required.

Once the optimisation model delivers schedules aligned with the company’s priorities, it is time to empower it to deal proactively with the impacts of disruptions by creating a significant degree of automation and intelligence. To this end, the optimisation model will constantly predict the occurrence of disruptive events and, if triggered, reschedule autonomously, preventing planners from being caught off guard and with outdated schedules. To accomplish this, it is essential to identify which events actually cause deviations and constantly monitor and analyse the relevant data points, such as supply delays and absenteeism.

Future-proof your design by strategically planning the system architecture for the production process, considering your aspirations for the end-state. It is critical from the very beginning for companies to envision the eventual goal and design their Next Gen production scheduling capability in order to decide on the details of the future system architecture and integration design (e.g., information flow and data model, etc.).

This makes possible an efficient, seamless transition from design to deployment and ensures that future business needs will be met by IT functionalities and a redesign will not be needed.

To finally unlock the potential of the Next Gen production scheduling and increase drastically the level of visibility and collaboration within the entire organisation, multiple country plants should be integrated for cross-supply visibility. This allows for greater scalability of the network.