Printer Friendly

Using computer simulation in the management of missile production programs.

Using Computer Simulation in the Management of Missile Production Programs

Department of Defense weapon system program managers must make decisions affecting broad areas of design, manufacturing and testing of their systems. A manager's performance is generally evaluated based on the cost, schedule, and performance of the system. Quality is also a growing concern for program management as systems become more complex in their design, manufacture, and maintenance. Often, program managers are forced to make important decisions based on qualitative assessments by subordinates, consultants, or the manufacturer, rather than on quantitative analyses. There is clearly an opportunity to provide program managers with quantitative tools for decision making.

The Analytic Sciences Corporation (TASC) decided to research new methods of providing missile manufacturing managers with timely information in support of:

* Program monitoring.

* "What-if" analyses required of missile program managers by budgetary and operational authorities.

* Decision making within the program office.

The frequent "what-if" analyses which program managers must perform center on changes in quantities, budgetary considerations (i.e. a stretch-out or compression of the program) and possible effects on cost and schedule of introducing a second prime contractor. The result of this research was to create a simulation based decision support tool for missile production. This article focuses on the effects of changing the buy strategy for a program of fixed size and duration.

There are a wide variety of missiles currently in production and operational in the field. There are also several which are completing full scale

development (FSD) and will soon be entering production. Missile systems range in size from hand-held units such as the anti-aircraft Stinger to submarine-launched strategic missiles such as Trident. Missiles launched from aircraft are also important, such as the Sidewinder air-to-air missile and ALCM cruise missile. Despite the variety and size of missile systems, TASC found that there is a high degree of commonality in the conceptual design and manufacturing processes of missiles.

Current weapon system programs use a variety of contractual requirements to receive data on the design, quality, and manufacturing progress of a system. Two of the most widely known requirements are MIL-Q-9858A (Quality Program Requirements) and MIL-STD-1528C (Production Management). In addition to the MIL-STDs each program has its own unique contract data requirements lists (CDRLs). Large quantities of CDRL data are regularly delivered to program offices. Some CDRL data may also be available through on-line databases, if it is required in the production contract. At times the relationship between the government and the prime contractor is adversarial. There may be disputes regarding the proprietary nature of data. It is also not necessarily easy for the government to gain access to the manufacturing facility or process data. In this environment the program manager must still monitor the contractor for cost, schedule and performance.

Missile program managers have several specific, recurring problems which may be addressed through computer simulation. These problems include:

* Determining the effects on the program of changing lot order volumes.

* The effects of "speeding up" or "stretching out" the program.

* Forecasting future states of the manufacturing system based on the current state.

In dual-source acquisition programs there are further needs which may be addressed through simulation. These include:

* Determining whether or not a proposed new manufacturer can deliver finished systems on time and within cost.

* The need to make decisions regarding competitive contract awards.

Most present analyses rely on static models of the production and competition environment. PERT and PERT-like methods dominate the field. A dynamic tool for production analysis would provide program management with additional capabilities and situational awareness.

The first step in developing a missile simulation methodology was to perform an analysis of the common traits between various missile systems. Missiles were divided into subclasses as part of this analysis. The subclasses of missiles were: hand-held (such as the Stinger), tactical (such as the Standard), cruise (such as the Tomahawk) and strategic (such as the Trident). The common traits of the missile systems were derived from analyses of:

* Manufacturing processes.

* Component subsystems.

* Work breakdown structure (WBS, MIL-STD 881A) for missiles.

The results of the analyses were definitions of processes, components and subsystems common to all of the systems and to specific subclasses of missiles.

Using the results of these analyses, a production paradigm was then created for a generalized missile prime contractor. The paradigm concentrated on operations which were performed in the prime contractor's facility and dedicated to specific missile programs. The contractor was given a set of shops for in-house production or integration of "parts." The prime contractor also received "items" he purchased (i.e. subsystems such as electronics packages), government furnished equipment (GFE) and "raw materials" for in-house fabrication operations. A graphical presentation of the paradigm is presented in the "missile production model" (previous page). This paradigm was then coded into a simulation model using the SIMAN language.

Aggregate statistics were developed for each of the individual shops visited by the parts during the simulation. Each shop, assembly, test and quality control area was defined as having a set of "resources" which together represented the overall capability for the area to handle concurrent processing. A cost index was created to allocate cost based on shop time and quality. This resulted in both part and system cost indices. The simulation model was validated using data from missile programs previously supported by TASC and the delphi method was used for validation of simulation outputs.

Further work on direct applications proceeded by creating an example missile program and using the simulation model to "manufacture" each missile in the program. Data to drive the simulation was provided from TASC databases on previously supported programs. Even with the aggregation of shop times and resources, a total of 130 system random variables were required to drive the simulation. The annual lot orders for five different missile program buy patterns were investigated. The quantities for each lot are shown in Table 1. A total of 5,000 missiles were included in each production program. It should be noted that extensive tailoring and many more variables will be required for full implementation of the simulation for a current missile program.

Results of the simulation include variability, as well as point estimates, of output variables associated with the program. This allows a program manager to gain understanding of the range of effects of varying order quantities and changing levels of quality on cost and delivery times. The effects of varying buy strategies on the cost of scrap, rework and repair and the delivery schedule were of special interest.

Plots were prepared based on the output data. These included plots of:

* Lot delivery times for each version of the program.

* Scrap/rework/repair costs for each simulation run.

These plots provide program management with data required to address several of the problems mentioned above. The great advantage of these is that as the program matures and changes occur, the simulation can be updated to reflect the current status of production. At that time the graphical tools can then be regenerated and thereby be kept up to date.

Figure 2 presents the lot delivery times (measured in years after initial contract award) for each run of each buy pattern. It is apparent that in each case the final lot is delivered at approximately the same time, however there are noticeable differences in both the first lot and intermediate lot delivery times. Buy patterns 2 and 3 have the most noticeable differences in early lot delivery times. Buy pattern 4 provides the earliest delivery times for lots 6 and 7.

A regression of the lot delivery time data is shown in Figure 3. The equation is y = 1.1087 + 1 .3964x. While the correlation coefficient is .98, it is apparent that the choice of buy pattern can drastically affect lot delivery times. Again referring to Figure 2, however, it is apparent that the equation is a fair predictor for buy patterns 1,4, and 5, especially with regard to lots 2 through 6. This information can be used to support management decisions on how a program of 5,000 missiles should be purchased in order to achieve required by the end user command.

The usefulness of simulation output is not limited to the determination of delivery times. The first pass inspection yields and probable dispositions of missile subcomponents and assemblies were put into the model. Figure 4 presents the mean unit costs of scrap, rework and repair for each simulation run. The scatter plot shows these variations. In particular the unit costs were low in runs 8, 10, 14 and 15. Repeated experimentation with a model can result in areas of possible cost and quality improvement being targeted. This could be in the form of process changes or reduction of the time and effort required to dispose of nonconforming materials. The final result would be a reduced unit cost of the missile.

This simple example illustrates how a simulation model may be used to help support fundamental decision making by a program manager. A full investigation would involve looking not only at a single total buy quantity for a program, but multiple buy quantities, varying total numbers of lots and the possibility of breaks in production. Other variables such as facility utilization, individual unit costs, queuing times and subcontractor dependencies would also be investigated. Together this can provide the program manager with a broad view of how a program may proceed, depending on actual or projected constraints.

Simulation can also provide missile program managers with a dynamic tool for analyzing the effects of ongoing program changes on their systems. This model can be used not only to look at changes in the dynamic environment, but it can also be used to generate graphical tools for rapid responses to inquiries. Further, the ability to examine the production system from a high level will improve program monitoring. This is achieved through using simulation output to validate contractor supplied data and highlight potential problem areas.

Further investigations will continue with the full implementation of the model for a current missile program. TASC is currently working with two programs toward full scale implementations of this simulation methodology and its use as a regular management tool. TASC is also continuing to invest internal research and development funds in the use of computer simulations in competitive evaluations and as a means of performing designed experiments. [Figures Omitted] [Tabular Data Omitted]

W.L. Scheller II is a member of the technical staff of TASC in Arlington, Virginia. Dr. Scheller is TASC's principal investigator in the area of simulation-based program management support tools. He received his BS in industrial and operations engineering from the University of Michigan, Ann Arbor in 1981 and his MS and Ph.D. in industrial engineering from the University of Nebraska in 1983 and 1989. Dr. Scheller is a senior member of IIE and SME and is currently treasurer of the Greater Washington Chapter of CASA/SME.
COPYRIGHT 1990 Institute of Industrial Engineers, Inc. (IIE)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Scheller, William L., III
Publication:Industrial Management
Date:Sep 1, 1990
Previous Article:Concurrent engineering, global competitiveness, and staying alive: an industrial management roundtable.
Next Article:Facts, creativity, teamwork, and rules: understanding leadership styles.

Related Articles
Modeling software tests design and system logic: real-time emulation and simulation are the keys.
Virtual foundry answers 'what if' production questions; computer simulations provided Foundry Service Co. solutions to its bottlenecks onscreen -...
Paperless factory plan helps jet fighter's takeoff.
Simulation Software Solution Does the Math.
Army pondering alternatives for tactical missile payload.
Smart missiles: weapons can switch targets in flight, send back pictures.
Missile defense agency prepares for key flight tests in 2005.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters