PPDS processes Dallmer relies on artificial intelligence for detailed production planning

Related Vendors

Abels & Kemmner Ges. für Unternehmensberatung mbH

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

The traditional company Dallmer relies on modern manufacturing techniques and a highly developed supply chain management. The latest addition is an AI-supported production planning and fine-tuning system, which replaces conventional planning tools.

Orhan Haydarlioglu is the ED manager at Dallmer.

Maik Babucke is in charge of the order center at Dallmer.

Dr. Bernd Reineke is the second managing partner of the consulting specialist Abels & Kemmner.

The company Dallmer, founded over 100 years ago, specializes in the development and manufacture of products in the field of drainage technology. These include floor drains, shower channels, drainage systems for terraces and balconies, as well as solutions for the use of rainwater. The manufacturing of the products, which are predominantly made of plastic and steel, is carried out using various manufacturing technologies, including injection molding technology and various mechanical manufacturing processes. In addition, assembly activities are carried out both by its own staff and by external service providers.

Dallmer has the highest demands on the functionality, design, and quality of its products. These demands also extend to the internal processes in the company-wide supply chain. A high readiness to deliver and timely delivery are at least as important to the company as quality. To ensure this, the manufacturer has been using optimization tools as an extension of the ERP functionalities in the area of planning and disposition for three years. Recently, an AI-supported PPDS (Production Planning and Detailed Scheduling) for production and sequence planning has been added. It was supposed to replace the old, aged fine-planning tool and thus increase both the quality of planning and the degree of automation.

The task of the new PPDS system is to allow an integrated consideration of quantities, capacities, and required resources and to achieve an optimal planning result in the process. Due to the high complexity with an insurmountable solution space, conventionally developed PPDS systems are not up to these requirements. Therefore, in everyday business, it is not uncommon for systems to only serve to visualize the current production plan, which was created based on comparatively few standardized parameters. However, not all existing restrictions are ever taken into account in this, because this cannot be calculated in a reasonable time with conventional mathematical methods. There is not yet the required computing power for this, and it will not be available in the foreseeable future. As a result, actual production almost always deviates from the target, and schedulers often have to rely on the expertise of production managers to somehow ensure the delivery readiness goals.

Diskover PPDS uses AI

However, Dallmer's claim is different: the goal is to generate a feasible production plan every week that already takes into account all existing restrictions in production and the availability of required resources such as machines, material, and personnel, and that can still be updated to the current day status at short notice in the event of new incidents. To be able to fully take into account the complexity of the production processes, artificial intelligence should be used. However, AI alone is not enough. Further essential functionalities were needed to make AI usable. Essential components are for example:

• Score functions for evaluating the solution scenarios and for mapping the company-specific target variables such as deadline compliance, productivity, and throughput,

• a working plan model for accurate and flexible representation of production processes

• and a clear presentation of the planning result in the form of an interactive Gantt chart.

The selected AI-based PPDS system from SCT Supply Chain Technologies offers all these functions and is specialized in finding the optimum in a seemingly infinite solution space. The medium-sized company Dallmer also already has enormous complexity in its production processes. During the optimization, the AI algorithms recognize which changes, so-called moves, lead to an improvement and which do not. All changes to the production plan are evaluated by the score functions and checked whether an improvement has occurred. If this is the case, the predecessor solution is discarded and the current solution is kept as the optimum from this point until an even better solution is found. The AI is able to anticipate the next steps and identify areas that can potentially be omitted. As a result, a multitude of possible solutions are not considered, but only the promising ones.

Deadline compliance comes first

At Dallmer, when setting the score functions, the greatest weight is placed on adhering to deadlines in order to ensure a high level of customer satisfaction. The second most important criterion is productivity, which is reflected in the avoidance of setup processes and machine downtimes. As in many companies, it also happens at Dallmer that a particularly important order should be given precedence over other orders. This can be achieved through so-called external priorities, which are taken into account in the score calculation and lead to a better result when giving preference to such orders.

The representation of work processes at Dallmer could not be based on the classic work plan model as it is usually implemented in ERP systems. Of particular relevance was the representation of the setup processes in the plastic area, as only specially trained personnel can be used for this. This setup personnel must be particularly taken into account in the optimization as a possible bottleneck. Therefore, an individually developed work plan model for Dallmer is used in the new AI-supported PPDS system, in which operations can be divided into any number of operation sections, referred to as tasks from here on out. In this way, all processes can be displayed and any resources such as setup personnel, material or tools can be assigned to each task. If alternative resources are available, they can also be defined per task without any complications. This now makes a comprehensive check on the availability of the required resources possible for the first time, and the created production plan can be implemented in all aspects.

The Gantt chart creates transparency

The interactive Gantt chart provides a quick overview of the planning result. The sequence of the individual tasks is shown for each workstation. In addition, the dependencies of the tasks on each other are visualized as a graphical or tabular network structure. These network structures also highlight problem areas, for example if a late material order prevents a task from starting on time. If the production scheduler wants to make a manual change to the plan, he can simply do this via a context menu. Not only are all possible placements of the task shown to him, but also the assessment of the shift using the score functions after the change. In addition, the production scheduler receives the list of all other tasks that were changed by his move. With the integrated undo function, he can easily revoke his intervention.

Diskover helps in times of changing bottlenecks

In total, Dallmer was able to achieve excellent results in fine planning within just a few months and considerably increase on-time delivery. "Especially in times of changing bottlenecks, where suppliers can't deliver and staff are out at short notice, the new AI-supported PPDS system has become an indispensable tool. It delivers better results faster despite increased complexity," explains Maik Babucke, head of the order center at Dallmer. The users have come to appreciate the system in a short time and wouldn't want to miss it. In a next step, the external service providers are also supposed to be controlled via this tool. By the way, Dallmer decided on the PPDS system from SCT Supply Chain Technologies based on the recommendations of the management consultancy Abels & Kemmner, who also accompanied the introduction.