Service Characteristics

.. re usually conventional units of time, most frequently weeks, although one-day, two-week, and one-month time buckets are not uncommon. The time buckets need not be the same for the entire schedule. Each product has its own MPS, and there is also an overall master schedule that synthesizes the requirements for ll products or a group of products that share facilities. The bill-of-materials (BOM) file lists for each end product all assemblies, subassemblies, components, and raw materials necessary to produce the product.

A BOM file includes information about how many units of each item are needed for each higher-level item in the product hierarchy (and possibly for the end product itself), whether the item produced internally or purchased, and the production or purchase lead-time necessary to acquire the item. A good way to visualize the hierarchical structure of the product is to use a product structure tree. The BOM is then used to construct a material list, which combines and summarizes all the material needs for the product. The inventory record file is a file listing the current inventories and outstanding purchase and production orders for each item. Although accuracy is important throughout the MRP system, the file whose accuracy is most crucial, and is most prone to error, is the inventory records file. A simple heuristic method that dos this is the part-period balancing (PPB) method developed by DeMatteis.

We first define a part period as a unit of measure that is equivalent to carrying one unit of an item (a part) in inventory for one period. Th PPB heuristic is based on the observation that in the basic EOQ model the optimal order quantity or lot size occurs when the total ordering or setup cost equals the total holding cost. Using this idea, the PPB algorithm first computes the economic part period (EPP) EPP = (setup or ordering cost) /(holding cost/unit period) An approach that does guarantee an optimal solution is the Wagner-Whitin algorithm. Wagner and Whitin formulated the lot-size and scheduling problem as a dynamic program that can be solved to find the optimal lot sizes over time. A dynamic program is a constrained optimization problem formulated in such a way that it can be solved using a special recursive (repetitive) technique.

Specifically, the problem is broken into decision stages that correspond to time periods. The algorithm sequentially determines the best action to take in the last (first) time period (in this case, whether an order should be released and, if so, the lot size to use), then the best action to take in the last (first) two periods combined, the last (first) three periods and so forth. The solutions at each stage are optimal, and the method used to solve the problem at each state is moderately efficient, although much less efficient than linear programming. Because of the nature of dynamic programming, multiperiod problems such as the MRP lot-sizing problem are especially suitable for solution in this manner. Although Wagner and Whitin first proposed their method in 1958, and although it provides an optimal solution for the given parameter values, it has not been widely used in practice.

There may be two reasons. First, dynamic programming is far less widely known by managers, engineers, and computer scientists than other operations management modeling and solution methods. Second, for even moderate-sized problems, obtaining the optimal solution is computationally slow and cumbersome compared to other optimization models and algorithms. A net-change system is one in which production and ordering plans are continually revised whenever the information on orders, production levels, and receipts is available. For example, if a customer asks to postpone delivery of a product by one week, the production and material requirement plans would be immediately revised to reflect this change.

Rather than reproducing all the plans, however we would identify only the changes to the plan and disseminate them through a change report. With a regenerative system information about order changes, material receipts, and actual production levels is gathered for a time period (say, a week). Then the MRP system is rerun, incorporating this new information and new material requirement plans and production schedules are generated. This type of updating is easier to manage because it can be done at the same time in every period. It is also less costly and creates a more stable operating atmosphere.

The only risk is that if major changes occur (e.g. a large order is canceled), the delay in updating the production plans can be harmful. One way to handle this potential problem is to allow for emergency updates of the regenerative MRP system. Time fences are periods of time during which changes to production and procurement plans are restricted. For example, a company may use two time fences: three weeks and six weeks. During the thee-week time fence the production schedule is essentially frozen, with only minor changes allowed: from three to six weeks into the future, larger changes in production and possibly resequencing of production runs may be allowed; beyond six weeks, unlimited changes may be allowed.

Like aggregate production planning, MRP uses a rolling horizon approach. That is, although a 15-week plan may be derived, only the first 1 or 2 weeks are implemented as planned. The plans for subsequent periods are revised and updated over time, but taking into account what is best over entire time horizon. A far more efficient way to address variations in lead times (including production lead times) is to use safety time, that is, to place orders or schedule production earlier than necessary so that if there is a delay the items will still arrive when needed. Because the production planning aspects of MRP ore related to most other functions of a company, the scope of MRRP has been expanded in recent years to integrate MRP with the order processing, billing, shop floor scheduling, and personnel and machine utilization activities of the company.

These newer systems, called manufacturing resources planning or MRP II, contain the classical MRP scheduling function as their centerpiece. However, MRP II systems may include a module that collects sales and customer order data and generates an MPS for future end product requirements (e.g., using a forecasting model). In addition, an MRP II system may convert information from the material requirements plans into specific work schedules for departments and machine, evaluate department workloads and capacity conditions, generate shipping documents and customer invoices, and produce management reports on production and financial performance. The benefits of using an MRP system 1. Low inventory Levels, Especially for In-Process Materials.

Because materials are acquired or produced only when needed and in the quantities required, inventories are kept to a minimum. 2. Good Material Tracking and Production Scheduling. The material requirements plans for each item provide a quick summary of the status of each item used in production: how much is on hand, the status of outstanding orders, and the schedule for production. 3.

A Method to Evaluate and Allocate Production Capacity. Tentative material requirement plans identify possible production bottlenecks and capacity problems. These plans can be used to decide whether to expand short-term capacity or reschedule production and how to reallocate production among time periods to stay within capacity limits. Little JIT is a form of production scheduling and control whereby 1. Items are produced only to satisfy actual demand 2. Production is performed in small lot sizes 3. Production is pulled through the production system by the last production stage rather than pushed through by the first production stage.

The consequences of Little JIT are smaller inventories and shorter throughput times Bit JIT (or lean production) is a complete reengineering of the production process that emphasizes continuous improvement, quality management, reduced setup times, improved maintenance procedure, and cooperation with suppliers. Two ways in which companies adapt to uncertainties in product demand are to 1. Improve their demand forecasting 2. Produce in anticipation of demand, that is, maintain final product inventories. The first solution is never perfect and can still result in stockouts or late deliveries, and the second solution increases costs. Other deviations from the ideal world are accommodated by doing the following 1.

Increasing inventories of raw materials to accommodate variations in delivery times and to take advantage of economies of scale in purchasing. 2. Increasing raw material, in-process and final product inventories, as well as overpurchasing materials and overproducing products, to make up for lost production due to defective raw materials and processing. 3. Scheduling large production runs and consequently holding large cycling inventories, to spread the fixed costs of machine setups over a larger number of units. 4.

Maintaining large in-process inventories between production stages to keep operations running during product changeovers at other production stages and to protect against delays resulting from machine breakdowns and employee absences. Production in classical systems is normally based on speculative demand. That is company forecasts the demand that is likely to occur in the future. Materials are then ordered, and the first step in the production process is scheduled. Subsequent steps in the production process are also scheduled, but their execution will depend on when earlier production stages are completed in other words, production is initiated by scheduling the first production stage, and the output is pushed through the production system. This approach is commonly characterized as a push production system.

Production of a specific product is initiated by the final stage of production in response to actual or assured demand, and production is pulled through the system, the final stage pulls the needed materials from the preceding stage, which pulls from its preceding stage, and so on. As a result, JIT scheduling has been characterized as a pull production system. Kanban (the Japanese word for card or ticket) to department 3. Department 3 puts that ticket in line for production and begins production on it within hours of receipt. The number and size of kanbans Several methods have been proposed for determining the best lot sizes (also called kanban sizes) and number of kanbans to use in a JIT system.

In practice, the size of a kanbans is determined by the time and cost of a setup, the dead rate, and the number of units that can be conveniently stored and transported. In 1950, while working for Toyo Kogos Mazda plant, Shigeo Shingo began his systematic study of production setups. Shingo found that similar problems occurred frequently during setups. At that point he made the key discovery that would lead to faster setups: setup operations actually consisted of two different activities: what Shingo calls internal and external. Internal setup activities, such as mounting a die or changing the in cartridge on a photocopier, are those that can be performed only while the equipment is stopped. External setup activities, such as collecting all bolts needed to mount a die or unpacking an ink cartridge from its carton, are those that can be done while the equipment is operating.

Shingos two observations formed the foundation of a procedure for reducing setup times that he called single-minute exchange of dies (SMED) (named for consulting projects in which he reduced the time for die changes on large presses from several hours down to less than 10 minutes). SMED uses the following four-step procedure 1. Observe and analyze how the setup is currently performed 2. Separate internal from external setup activities 3. Convert internal to external setup activities. Cleaning activities can often be transferred from internal to external setup. For example, by having two sets of tools or two processing vessels one can replace the dirty or contaminated one with a clean one quickly during internal setup, and then the dirty one can be cleaned after production has restarted.

For fluid vessel cleaning rather than having two separate vessels a lower-cost solution is often to use plastic vessel liners that can be exchanged quickly. 4. Simplity and streamline activities (a) Provide each work station with its own tools and store the tools conveniently (b) Standardize the size and shape of dies and other parts that must be changed during setups (c) Use the same fasteners, such as bolts, for each setup. The fasteners can stay with the machine they simply have to be loosened ad retightened for each setup, rather than removed, replaced, transported, and stored. (d) Use fasteners that can be loosened and tightened with a single turn rather than those that require turning the fastener several revolutions. (e) Reduce or eliminate adjustments by using fixed settings and markings on dies, tables, and guide bars.

There are at least four ways in which customers and suppliers can work together to provide reliable delivery to the customer at lower cost to both. 1. Share production scheduling plans quickly 2. Include suppliers in product design 3. Help suppliers improve their production methods 4.

Have spatially close facilities Causes of machine failures 1. Inadequate preventive maintenance 2. Overusing and operating machines at excessive speeds 3. Dirt, oil, and chemical damage 4. Collisions (e.g. a fork lift hitting a machine) 5.

Incorrect machine setup for operation 6. Materials fed into machine or processed incorrectly. The cost of machine breakdowns can be substantial, yet eliminating these six causes can be relatively inexpensive. That is the focus of the following maintenance strategy, which is sometimes called total productive maintenance (TPM) More Maintenance Principles Several other principles and actions can enhance equipment maintenance 1. As in product and process design, the secret of better machine maintenance is simplicity 2. To solve maintenance problems and to allocate maintenance resources efficiently it is important to collect useful data on the frequency and causes of machine failures 3. When some parts of a machine start to wear out, many other components are wearing out as well.

It is better to rebuild the machine by replacing all worn parts at once instead or replacing them one by one, as it is sometimes done. 4. It is surprising how much time is wasted when a machine is shut down for overhaul and key components are not available when needed. Procurement should be carefully planned so that all parts are on hand when needed. Institutionalizing this principle of continuous improvement (or kaizen in Japanese) is an integral part of any Big JIT production system. We have already seen several specific ways to do this SMED, TPM, and TQM are all way to generate continuous process improvement.

The most successful continuous improvement systems are those that encourage employees themselves to find ways to improve production methods and to improve products. Three ways to promote this process are to use employee suggestion systems, quality circles, or autonomous work teams. Employee suggestion systems have been used for over a century, and some companies have reported astounding success with them. Quality circles are small groups of employees who met regularly to discuss and evaluate ways to improve productivity, quality, and safety. Autonomous work teams are groups of employees who work together as a tam to perform some aspect of production.

They are often evaluated and rewarded as a tram, and one aspect of the evaluation is improvement in performance. The development of both quality circles ad autonomous work teams was motivated by the continuous improvement philosophy. JIT production When to use JIT scheduling The factors that most significantly affect the desirability of using JIT scheduling are the dead pattern, the product mix, and the structure of the production process. 1. Be patient 2. Customize implementation 3.

Be flexible and adaptive 4. Maintain excess capacity Synchronous (synchronized) production is a procedure for balancing the flow of production. It focuses on keeping the bottleneck stages fully utilized and forcing the other stages to produce in synch with the bottleneck. The mechanism for forcing the stages to produce at the same pace is called the drum-buffer-rope procedure. Another approach to production scheduling in inventory management that is especially useful in job-shop environments is synchronous (or synchronized) production.

Synchronous production has some similarities for JIT production, such as recognizing the harm resulting from production variations, trying to coordinate and pace production so all production stags are producing at the same rate, and using small batch sizes. However, synchronous production differs from JIT in two ways. 1. It focuses more on adapting efficiency to variations and imbalances in the system, rather than eliminating them 2. Production scheduling is hybrid of classical push and JIT pull approaches. Synchronous production, developed by Eliyahu Goldratt, is based on the theory of constraints.

This theory was developed from three empirical observations. 1. In multistage production systems not all stages have the same production capacity 2. Variations are randomness in production systems reduce effective capacity and output 3. The procedures used in classical production systems generally amplify rather than solve the problems created by capacity imbalance and production variations.

A bottleneck resource is any resource that limits the flow of production through the system. A nonbottleneck resource is one hat has a capacity grater than the demand and therefore does not redistrict the system throughput. A capacity-constrained resource is not a bottleneck but is operating close to capacity and could become a bottleneck if not operated efficiently These facts lead to the key principle of the theory of constraints, to manage production effectively, one must focus on the constraining resources the bottlenecks. Goldratte. His key recommendation is a mechanism for coordinating production so that the flow of production (rather than the nameplate capacities) among the stages is balanced, and inventories are kept to a minimum except where they are really needed.

This scheduling method is called the drum-buffer-rope method. The drum is the mechanism that controls the pace of production; that is, the drum beats the rhythm of production. In front of bottleneck operations a supply of safety stock is maintain to buffer or protect the bottleneck operation from material shortages, the bottlenecks must be kept operating. The rope is the link between the bottleneck and preceding work stations that keeps them from running ahead of the bottleneck. In JIT scheduling the rope is the kanban pull mechanism. In synchronous production the rope can also be kanbans pulling production, or flow can be controlled by using a daily production schedule based on the production of the bottleneck.

JIT, synchronous production and MRP All three can be integrate to varying degrees, depending on the situation. Synchronous production fits well with a JIT production system that has some imbalances and variations, which most do. The drum-buffer-rope method of controlling production flow is especially helpful in job-shop processes, where JIT scheduling may be impractical due to the variety of products and product routings. One subsystem may best be driven with a JIT mechanism, while another may b better suited to MPR. (In fact, these three philosophies use different approaches to accomplish the same goals: maintain low inventories, deliver products on time, and avoid stockouts.

Words 6.319.