Simulation Supports Production Planning and Scheduling

Almost all large, and the majority of small and medium-sized enterprises (SMEs) run MRP / ERP systems. However, many are not satisfied with their system. This paper investigates the reasons for dissatisfaction. It points out some weaknesses of conventional MRP tools and explains how advanced planning and scheduling tools can be used. A number of examples illustrate different ways to use simulation techniques for an improvement of planning tools and methods. 1. Conventional Production Planning Conventional production planning and scheduling is done in a top-down way, with decreasing planning horizon (from years to hours) and increasing level of detail. Expected or real orders are allocated to production resources in such a way that the overall objectives: • maximal capacity utilisation • minimal throughput times • minimal delays • minimal WiP are reached. However, these objectives are partly in conflict with each other, hence every user company has to define their relative weight. A major problem of conventional production planning is the lack of information about the situation on the shop floor: Actual state of resources and their work load are not known. It is assumed that all resources are available. On the other hand, process times are often increased by “transition times” to cover the loss of time during transportation and waiting for free capacities. Transition times –which in reality depend on the actual situation – are fixed data and based on experience or negotiations with shop floor staff. Production plans generated on this basis are often unrealistic, and often the foremen on the shop floor ignore them and do their own planning. Since the transition times in use are often extremely long (see table 1 below), delivery dates which are calculated on their basis are often not accepted by customers. Sales staff therefore needs to agree shorter dates – which leads to increased chaos on the shop floor. Simulation Supports Production Planning and Scheduling Page 2 of 13 2/13 2. A Case Study A medium sized manufacturer of toys in Germany runs an MRP system since many years, but they are not satisfied with it (/1/). An analysis of data stored in the system and of applied planning procedures revealed that: the system assumes 140 work stations but allocates orders to only 72 of them. 52 stations are ignored, 14 stations do not (no longer) exist in reality. the assumed throughput times per department wee agreed with the respective foremen some time ago and deviate substantially from the process times as shown in table 1. the planned utilisation of some work stations is 500%! feedback from shop floor to the MRP tool is only provided on weekends, hence orders are often released much too late. many orders have a delay of several months warehouse contents in the MRP system and in reality do not coincide at all. Thus often orders are released even though the material is not available. the department heads on the shop floor do their own scheduling using Excel. However, these schedules must be changed very frequently. As a result, many potential customer orders are lost, on the other hand many products are erroneously produced to stock. /1/ Department No. Assumed Throughput Time Real Process Time 1 22 labour days 1 hour 2 5 labour days 22 hours 3 7 labour days 1,5 hours 4 7 labour days 10 hours 5 15 labour days 1,5 hours 6 15 labour days 20 hours 7 5 labour days 1,5 hours Table 1: Assumed throughput times and real process times of departments (/1/) 3. The Contribution of Simulation Technology Simulation is well known as a powerful tool supporting the design, layout or re-design of factories and production systems /2/. Recently, many successful applications proved that it can also support the operation of manufacturing systems, especially in the area of scheduling and control. Experience documented below proves that also SMEs can take advantage of these developments. In general, one can distinguish two ways of using simulation for improving production planning and scheduling: Simulation Supports Production Planning and Scheduling Page 3 of 13 3/13 1. A simulation model is used to configure, test and fine-tune an existing planning tool. Many planning tools offer a wide range of parameters and option to allow their adaptation to different application areas. However, often the number of possible configurations is so large that even experts cannot predict any more which is best or even good. A simulation model can be used to test configurations and evaluate their effects on the company’s objectives. This way of testing is much faster and less expensive than testing in reality. 2. A simulation model is used “online” as part of the planning tool box and rund in parallel to the real production process. It can take into account all kinds of rules and constraints and does the bulk of routine work for the human planner. Using real shop floor data, it calculates exactly the throughput times and can very quickly evaluate the effects of schedule changes. These tools are known as “Leitstand” or “APS (advanced planning and scheduling) tools. Most of them have a built-in optimisation algorithm which automatically schedules orders in such a way that enterprise objectives are met in a very good way. These two ways of using simulation are not to be seen as alternatives. Sometimes it makes sense to use both of them. This is illustrated in example 4.3 below. 4. Practical Examples For the Use of Simulation in Support of Production Scheduling First, three examples illustrate the first way, then three examples report about the benefits of introducing simulation based APS tools. 4.1 Case Study Continued The enterprise mentioned in the case study above used simulation in the first way and found that substantially shorter throughput times can be achieved, and the extreme work load at many work stations would decrease automatically as a by-product the flow of production would be much smoother if special strategy for releasing orders were applied. This strategy should be oriented to balance the load of bottleneck stations. As a by-product, the need for short term rescheduling would decrease dramatically. These findings shall now be put to practice in the enterprise /1/. 4.2 Example: Manufacturing of Tools This company produces a large variety of cutting tools (/2/, /3/). The production process consists of eight steps with several tests in between. The company used Simulation Supports Production Planning and Scheduling Page 4 of 13 4/13 simulation in order to find ways how its profitability could be increased. Among others, also the use of the MRP system was to be investigated.. The results of simulation were very surprising: All activities that were taken into account by the management proved useless. Some examples: Optimisation of lot sizes: The lot sizes had already been optimised in earlier projects. More machines: Useless since bottlenecks were moving. Additional machines would not be utilised at an economic level. Concentration, rejection of small orders: in this company, small orders can be used as “fillers”. On the other hand, simulation experiments proved that by a new organisation and a more flexible way of production planning, throughput times could be reduced by 50% work in progress would decrease substantially delivery reliability could grow by 10% The necessary changes would not require any major investment, but a higher and more flexible qualification of workers more flexible shift models a radically new configuration of the MRP system so that it can exploit the increased flexibility of resources These measures are now being realised one by one in the company. The company expects saving of more than 500,000 . The expenses of the study were approx. 40,000 , and the total payback time of the re-organisation project is less than 6 months. 4.3 Example Decorpart DECORPART is a medium-sized company, which supplies small, pressed aluminium parts to a range of other consumer-focused businesses. Typical applications include spray assemblies for perfumes and dispenser units for asthma sufferers. The business lies in a highly competitive sector and success depends on achieving high efficiency and low cost of manufacturing. In the past, Decorpart had already installed simulation based, finite capacity scheduling tools supporting the scheduling of individual areas of the production process. This had led to substantial increase of planning accuracy and allowed them to reduce the stock of raw material by 300,000 GBP (/4/). To improve the overall performance, increase output and reduce lead time, they now planned to implement an overall scheduling system co-ordinating all local systems. Modern production scheduling tools are very powerful and offer a range of options and parameters for adapting the tool’s behaviour to the requirements of the real process. However, the more options exist, the more difficult it becomes to find the Simulation Supports Production Planning and Scheduling Page 5 of 13 5/13 best configuration of the tool. Even experts often cannot predict the effects of the many possibilities. Testing out even a small number of possible configurations in reality, and studying their effects on the real production process might take months and might severely reduce the overall performance. Hence such tests are not feasible in practice. In order to deliver their customer the best possible solution, the supplier of the scheduling tool, Preactor International decided to use a simulation model for finding the optimal configuration of the scheduling tool (5). A custom-built model was built that simulates the arrival of orders, their queuing and their flow through all steps of the production process. For the overall co-ordination and schedule optimisation, each process stage was modelled as a group of machines with an overall capacity per day or per week. In order to check if the model reflects the real process adequately, a set of real data was compared with the data produced by the simulation. It was found that the model and the real process produce more or less identical results. The anodising process was known to be particularly important for the overall production. Therefore the model of this process was refined and the individual anodising tanks were described in detail, so that colour changeover and set-up times could be studied more precisely. In this way, it was tested to what extent the overall lead time of orders can be reduced by optimisation of the anodising process stage. Next, the Preactor scheduling tool was coupled with: a high level manufacturing/business system model, a detailed representation of the anodising process, both of which were prepared by Riga Technical University. These two simulation models were used for testing initial configuration of the scheduler and for iterative optimisation of its parameters and rules off-line prior to its implementation at Decorpart. Simulation Supports Production Planning and Scheduling Page 6 of 13 6/13 A high level simulation model of Decorpart’s entire business/production The anodising process stage sub-model The high level model of the entire business provided the following results: Simulation Supports Production Planning and Scheduling Page 7 of 13 7/13 1. If the response time for customer enquiries could be reduced by 5 %, the total revenue of the company would grow by about 10 %. The maximum revenue could be achieved if c planning time would not exceed 6 minutes. 2. By introducing the automatic PREACTOR Supply Chain Server, a maximum response time of 6 minutes per inquiry can be achieved. 3. In this case, the number of cancelled orders can be decreased by 14-18%, which would cause the total revenue or value of confirmed orders to increase by 100% (or twice). 4. Instead of four planners, only one would be needed if the PREACTOR tool were introduced. Thus, employment cost of approx. 150 000 Euro per year can be saved. The detailed model of the anodising stage led to the conclusion that improved sequencing rules for incoming orders in a week could reduce the total lead time of this stage by at least 4 hours, in some cases even by 19 hours. As a result, the production rate of the anodising stage will go up by 10%, and a significant increase in equipment utilisation and reduction of unit manufacturing cost can be achieved. Decorpart will save an amount equivalent to the total project costs in less than three months!! This project once again proves that often little investment is needed to test and optimise an operating planning too by simulation. Model based analysis can help discover enormous potentials for improvements and saving. 4.4 Example: Simulation Based Scheduling of Car Painting Today nearly every car in automotive production is different because of specific customer requirements. In the production step of painting, more than 100 different colours are possible. Combined with different car shapes like standard or station wagon, engine varieties and other options, the problem of finding optimal product sequences is extremely difficult. In contrast to the assembling process, where tools and process steps are very similar and uncritical in sequence, the large number of colours is critical: Each colour change requires a cleaning operation consuming considerable extra time and cleaning supplies. The cleaning process is easier and faster if the colours are changing all time from a bright to a darker colour. The buffers in front of and behind the painting station are limited in space. Therefore an optimised sequence of painting operations must be calculated, which meets the buffer restrictions and minimises the consumption of resources and time due to colour changes. An improved batch sequence can lead to major savings in costs and time (/6/). A graphical process model was developed. It was very helpful in internal discussions among the manufacturer’s staff, and in discussions with the external simulation Simulation Supports Production Planning and Scheduling Page 8 of 13 8/13 experts. The validated model was integrated with other IT systems so that the model always reflects the actual state of the real process. To generate and optimise schedules which satisfy all constraints, an intelligent optimisation toolkit ISSOP was then coupled with the simulation model as shown in the diagram below: The optimiser ISSOP generates production sequences, the simulation tool uses them as input, it checks if all constraints are met and calculates the costs and throughput time. ISSOP then compares the results and generates better sequences, the simulation model evaluates them again, and so on. After some iterations, which take between 5 and 25 minutes, an improved sequence has been found which is then used for controlling the real process. This approach increases the reliability of planning, reduces material consumption, throughput times and work in progress. Additional advantages of the simulation model are the possibility of testing the schedule against disturbances such as technical problems or delivery delays.