TD1G1Contrasting TOC and MRP Using Spreadsheets and Simulation,

C. David Wieters, Department of Management, College of Business Administration and Economics, New Mexico State University, Las Cruces, New Mexico,

The Avraham Y. Goldratt Institute simulations dramatically show the intuitive advantages of Theory of Constraints (TOC) relative to MRP/CRP. The five focusing steps of TOC are enough for students to out perform a disciplined application of a MRP schedule. MRP without CRP would have provided virtually no guidance within the simulation’s daily time buckets. Excel spreadsheets were used to create a MPS with CRP loads within daily capacities of all resources. The schedule limitations of MRP are highlighted. Without MRP, students focused on the CCR and profited. MRP planned order releases consistently led to significantly lower throughput.

INTRODUCTION

In an introductory Production Planning and Control course it is easy to become narrowly focused on procedures such as MRP logic, JIT shop floor techniques, or scheduling rules. This approach can easily detract from learning and understanding how to deal with operating issues and decisions. Insights into the complexities of job shop operations can be effectively revealed by combining a spreadsheet approach to MRP logic and the dynamics provided by the Avraham Y. Goldratt Institute Jonah Course simulations (Anonymous, 1992). This paper describes how MRP and CRP analysis, using spreadsheets, can be applied to Goldratt's SIM10. Executing SIM10 based on MRP/CRP guidance is contrasted to student execution guided by Goldratt's five focusing principles.

SIMULATION ENVIRONMENT

The Goldratt Institute has developed an extensive set of educational simulations running on a simple DOS platform. Sim10 was use in this teaching innovation. Sim10 includes three end products, two merging assembly points, one diversion point and 19 operations in total. There is a bottleneck resource. The simulations allowed students to learn by discovery. Teams of three learned the mechanics of the simulations on Sim30, a non-bottleneck model. These teams then worked Sim10 without guidance. Initial runs resulted in poor throughput or bankruptcy, partly due to misunderstanding the simulation, but largely due to activating resources rather than utilizing them. They did come to identify the bottleneck resource. Analysis of load versus capacity of each resource further supported their intuitive conclusions. Prior to further simulation runs, we studied the five focusing principles of Theory of Constraints.

THEORY OF CONSTRAINTS FOCUSING PRINCIPLES

Goldratt and Cox (1986:307) popularized the five focusing principles that lead us through problem identification and resolution. The Simulation:MRP/CRP experience addresses the first three principles: (1)Identify bottleneck, (2) Exploit the bottleneck, and (3) Subordinate non-bottleneck resources. Principle (4), Elevating the bottleneck to higher capacity and (5) Avoid Inertia by seeking the new bottleneck were beyond the scope of this innovation. The Identify step is achieved intuitively by student experiences in their initial unguided simulation run. This is followed by calculating load versus capacity in a rough-cut capacity planning in a Capacity Bills fashion (Vollmann et al. 1997:126-8).

The Exploit step is achieved by setting priorities based on throughput per minute of the bottleneck. Thus, the weekly production goals reflect a throughput maximization criterion. Foregone markets are planned for the products that generate less throughput per minute of bottleneck resource. The bottleneck schedule is based on ability to feed the operations on this key resource. Current work-in-process and required process flow times determined the ability to supply the bottleneck.

The Subordinate step was implemented at an intuitive level. While a more formal approach might add value, the informal approach does highlight the contribution of focusing efforts on the bottleneck and avoiding dilution of attention across all resources. The quickest successful teams used a "flocking" approach. This strategy focused all relevant resources in supplying the first schedule job on the bottleneck. Once this supply was achieved, the "flock" of subordinate resources moved to the next supply need of the bottleneck. As they got ahead, resources were diverted to producing products that were market, rather than manufacturing constrained. In a more "balanced" plant this "flocking" behavior risks avoidable lost throughput due to exceeding the "protective capacity" of the non-bottleneck resources (Shrikanth and Umble, 1997:101). However, in Sim10 simplicity is nicely rewarded. On their second run, three teams achieved their throughput goal (95% of theoretical maximum), a fourth improved significantly, and the fifth orally reported goal achievement.

At this point the class has studied MRP and CRP. The next step, and the key contribution of this paper, is to see if MRP/CRP approach could match or outperform the student's intuitive application of TOC concepts.

SPREADSHEET MRP-CRP ANALYSIS

Sim10 is displayed as a process flow chart consisting of 19 nodes, representing operations each color-coded for the resource required. Each operation yields a unique part. One of the parts is common to two end products, thus is common to each of their bill of materials. There are two assembly or merging points demonstrating the explosion of planned order releases to more than one component. Setup times differ for each operation depending upon the equipment required. The setup times critically affect the production plan, as does the presence of work-in-progress inventory in front of the bottleneck. Without setups, lot sizing would not be an issue. Without work-in-progress, the students would not see the utility of buffers, nor planning a position for the next planning period. This level of complexity goes beyond the typical textbook MRP or CRP examples that can be handled in reasonable time manually.

A spreadsheet model was developed and used by the instructor to compute net requirements, to time phase these requirements, and to explode these needs based on proposed master schedules of the three end products. Each planned order release was used to compute a time phased loading on the associated resource, thus serving the CRP function. The spreadsheet was presented in class, but the students did not have an opportunity to apply it to Sim10.

Refining the MRP spreadsheet highlights the challenges faced in making MRP/CRP work. Resolving these problems offers teaching opportunities. The difficulties were inaccurate lead-times, lot sizing decisions and lot splitting. Each of these problems has their parallel in reality.

A daily time bucket was used to provide the detail needed to work within the two-week duration of the simulations. Even so, serious detail is lost in a time bucket approach. If a minimum lead-time of one day were assigned to each operation, our lead-time would exceed our planning horizon. Lot-for-lot and zero lead-times were used to allow spreading workload throughout the week. This results in the "lot size" of an end product becoming a building block for lot size of its components thus affecting bottleneck operations. While a limitation, this constraint does reflect MRP when lot sizing is not based on a bottleneck capacity.

The spreadsheet assists developing a feasible daily MPS, the resulting time phased material plans and a time-phased capacity requirements plan for the week. As in real MRP systems, there are assumptions (zero lead-time in our case) that deviate from reality. In practice, validity deteriorates between re-planning points. In our case the validity degrades primarily due to our lead-time assumption. In life, we add random disruptions. A disciplined execution of the resulting MRP plan for Sim10 dissolves early with resources completing non-critical jobs rather than break a setup to support the bottleneck. Once recognized, crisis management prevails. This is a realistic demonstration of what happens when a plan becomes invalid.

TOC suggests that capacity and market orders should determine production lots at the bottleneck. To reflect this in the spreadsheet we can introduce the equivalent of Firm Planned Orders at the bottleneck. The resulting material and capacity plans yield almost optimum Sim10 results. The plan still requires early starts of some jobs on non-bottlenecks to eliminate temporary overloads. The Firm Planned Orders effectively become the tool to implement TOC logic where MRP and CRP can not. Firm Planned Orders amount to manual override of the MRP system, which we would like to minimize.

CONTRASTING OUTCOMES

In summary, TOC concepts that focus attention on bottlenecks and subordinates non-bottlenecks allow students to successfully manage the Sim10 environment. Conversely, the detail provided by MRP/CRP does not lead to comparable success without substantial manipulation to combat limitations in basic MRP and to exploit the bottleneck. This contrast reveals planning and control issues that would be overlooked in addressing TOC and MRP separately.

ACKNOWLEDGMENT

This project is a result of a 1993 Goldratt Educational Grant (Jonah Course) to New Mexico State University and the author. That support is greatly appreciated.

REFERENCES

Anonymous. Functional Educational Workshops . (software Ver. 4.1) New Haven, CT: Avraham Y. Goldratt Institute, 1992.

Goldratt, E., Jeff Cox. The Goal. (2nd Revised ed.) Croton-on-Hudson, NY: North River Press, 1992.

Vollmann, Thomas E., William L. Berry, D. Clay Whybark. Manufacturing Planning and Control Systems. (4th ed. ) New York: Irwin/McGraw-Hill, 1997.

Shrikanth, Mokshagundam L.and Michael Umble. Synchronous Management: Profit-Based Manufacturing for the 21st Century. (Vol.1) Guilford, CT: Spectrum Publishing, 1997.

Proceedings of the Tenth Annual Conference of the Production and Operations Management

Society, POM-00, March 20-23, 1999 Charleston, S.C.