Using Advanced Analytics to Boost Production Planning Optimisation 

In a dynamic environment characterised by rapidly changing needs, many companies see production planning as an unnecessary, slow and inefficient exercise. For most organisations, and regardless of their business sector, the difficulties faced by planning teams are similar; the complexity of communication between departments, the multiplicity of products in each order - produced in different manufacturing lines - and the lack of visibility regarding the impact of order allocation, are just some of the factors that constrain the planning process. Moreover, when unexpected events or changes in the strategy occur, the time spent creating and implementing a new production plan is perceived as a waste of resources.

Over the past decades, technological evolution has enabled the technical and organisational advancement of most companies. Production flows are increasingly more automated and restricted which culminates in increased complexity in the management and control of the entire chain. Given this scenario, the best planning teams are those that are able to master two key factors: the use of robust decision support tools and deep knowledge of the different planning models. Thus, it will be possible to control and establish plans that allow profitability of the productive activity, reducing setups and improving delivery.

What are the main layers of production planning?

The pull planning model, part of the KAIZEN™ methodology, has three action areas. All of them display a distinct degree of detail and aim to guide the entire chain according to demand, thus reducing waste.

  • High level - strategic planning: Decision, at a commercial and production level, of the references that should be kept in stock and that should be constantly replenished (MTS - Make To Stock), as opposed to those that should only be produced when orders are placed (MTO - Make To Order).
  • Medium level - capacity planning: Definition, at the production level, of the lines and shifts necessary to meet the proposed deadlines and targets, according to demand - for example, if demand in a particular week was irregularly higher, a decision is made to open more shifts or more production lines.
  • Low level - execution planning: Sequencing of operations based on what was determined in the medium level planning with allocation to the machine, and time when each production will start, following the sequence of operations and maximising efficiency.

How the layers of the pull planning model interact with each other

These three layers are closely related. The strategic layer, by defining the service level for each of the references, makes it possible to create a commitment, not only towards external customers, but also internally, i.e. towards the teams responsible for the fulfilment of the layers below. Similarly, the execution layer cannot start its sequencing work if, at the medium level, the necessary capacity to meet demand is not communicated.

It is vital to align the different planning models so that the three areas complement each other in a consistent and coherent way. In this context, the use of analytical tools facilitates the alignment between the different planning layers as it boosts the visibility of the production in real time. But how can planning models and data analytics leverage the technical knowledge of each flow?

The role of data analytics in the decision-support process

For each of these layers, decision-making support and work tools are needed. Traditional spreadsheets can quickly analyse simple data and information to determine the first steps of a pull strategy, categorising MTS and MTO references, and making weekly or monthly allocations according to installed capacity. The same spreadsheets can also transform this weekly planning into a production sequence, despite their limitations.

The increase of complexity and restrictions of the productive process, together with the massive data collection, enhanced by the Industry 4.0 models, promote and trigger the need for more complex planning tools. The answer is then to invest in allocation and optimisation solutions.

The importance of transversal and integrated visibility of operations

These new alternatives allow the production process to be modelled in a holistic way. The final goals to be optimised are clear, but they hide behind the complexity that truly limits manual planning: the vision and integrated assurance of compliance with production constraints.

One of the approaches is to use robust optimisation engines which allow us to achieve the optimal production sequence in the face of an established goal (for example, minimising the number of setups or the work in progress - WIP). Alternatively, and whenever faced with highly complex problems, these optimisation engines are replaced by a set of heuristics applied to the reality of each company and the characteristic constraints of each process.

With these tools, planning and supervision teams have access to a set of functionalities that allow them to monitor, predict and anticipate, truly contributing to the change of a sustained paradigm, ‘based on reaction’, to a reality in which they can plan in a more solid way and clearly assess the impacts of the reaction effects.

Features that allow you to adopt a new concept of operational visibility

  • Online recalculation - the ability of the system, upon the occurrence of any change, such as the entry of a new order or equipment breakdown to automatically recalculate a new solution to incorporate the new constraint.
  • Line blocking - for certain priority lines, the system allows the allocated production to be fixed, and is able to plan production with that block in mind.
  • Integration with existing ERP and MES systems - a complete integration with existing databases which dismisses manual imports and exports of information.
  • Business intelligence tools - the creation of dashboards that support decision-making, enabling the impact of actions to be verified and the production process to be controlled.
  • Visualisation and monitoring solutions - allowing visual control (with tools such as Gantt charts) of the production of each piece of equipment as well as the main process indicators.

What are the benefits of advanced analytics?

Online recalculation and integration with the different ERP and MES systems allow teams to be released for activities with greater added value, focusing on the discussion of various alternatives for prioritising orders to provide a better service level to customers.

Given the specificities of use, and complexity of the problems, several techniques and methodologies have been used. With the recent boom in the use of Artificial Intelligence (AI) and optimisation engines, it is possible to apply several strategies to solve these problems such as genetic algorithms or neural networks. Once the model has been created and validated by establishing its main variables and constraints, the algorithms have proven to be important allies of the planning teams regarding the necessary replanning and the determination of the new production scenario caused by the change of a certain factor.

These solutions, combined with the monitoring of dashboards updated in real-time, allow planners, supervisors and commercials to follow the production of the most critical orders and make data-based decisions, since many times, anticipating the production of a certain reference implies a reduction in the efficiency of other lines, generating more set-ups and stocks. BI tools enable the key indicators of each decision to be understood and monitored, allowing the management of priorities to be more fluid.

Finally, the rise of Industry 4.0 significantly boosts the entire data analysis component. The use of sensors, and the high connectivity between all the systems in a factory, allow not only problems to be detected more quickly, but also to react practically instantaneously. In particular, a failure detected by a sensor can automatically trigger the recalculation of a production plan, and notify the salespeople responsible for the orders affected by this failure.

Another case would be a quality defect, detected in a piece of equipment, which requires a reaction to be triggered - an additional number of parts to be produced that will naturally occupy the equipment for longer, and have an impact on the delivery of the final product.

The paradigm shift and the future of production planning

In conclusion, the production planning team paradigm is changing. The manual management of production plans without visibility of the impact of each decision, making the process of coordinating the various lines and variables time-consuming, has been replaced by automation with the implementation of advanced planning systems, that not only enable detailed analysis of each available production option, but also generate those same options and production sequences.

We are living a new reality in which analytical solutions go beyond the purest concept of automation, incorporating an intelligence capable of incorporating the constant challenges that mirror the volatility in demand, which therefore requires more holistic and efficiency-generating approaches throughout the chain.

Get all the latest news about Kaizen Institute. Subscribe now.

* required fields

arrow up