ÆäÀÌÁö À̹ÌÁö
PDF
ePub

N67-13766# Deutsche Forschungsanstalt fur Luft- und
Raumfahrt, Brunswick (West Germany).
NETWORK PLANNING TECHNIQUE. A SURVEY ON THE
METHODS OF CPM AND PERT [NETZPLANTECHNIK. EINE
UBERSICHT UBER DIE VERFAHREN CPM UND PERT]
A. Papendieck Mar. 1966 46 p refs In GERMAN; ENGLISH
summary

(DLR-MITT.-66-12) CFSTI: HC $2.00/MF $0.50

During the last years the field of time scheduling for large projects has seen the fast expansion of a new planning method know by many different names. The most frequently used names are Network Planning Technique. Network Analysis, PERT (Program Evaluation and Review Technique), and CPM (Critical Path Method). The number of different names has been estimated to approximately fifty. The methods, however, are only slightly different from each other, the essential principles are inherent to all of them. This report considers the main ideas. Finally, a computer program is described. Author

[blocks in formation]

N67-10520# Texas A&M Research Foundation, College Station.
SIMULATION OF PERT PROJECT COMPLETION TIMES
BY STRATIFIED SAMPLING METHODS Technical Report
No. 3

Larry J. Ringer [1966] 11 p refs
(Contract DA-31-124-ARO(D)-282)
(AROD-4721-4; AD-637830) CFSTI: HC $1.00/MF $0.50

The algorithms for computing the cumulative distribution of the completion times of PERT networks proposed by Hartley and Worthan (AD-626 486) and Ringer (AD-630 841) will in general only reduce the original network to a network with fewer activities. The c.d.f. of the project completion time for the reduced network will then be computed by some other method, such as Monte Carlo simulation. In this note sampling schemes for estimating this c.d.f. are discussed. A completely random sampling scheme and a sampling procedure based on stratification of activity completion times are described and compared. Author (TAB)

[blocks in formation]

the test cases; and these cases are considered according to user communities, interdisciplinary bias of the communities, information input types and their transformation within the system, processes to carry out the transformations, and cost factors related to information types and processes. The presented methodology is considered to have a high degree of usefulness for testing information system configurations and for general management planning. M.W.R.

N66-80779

Operations Research, Inc., Silver Spring, Md. A GENERALIZED NETWORK APPROACH TO THE PLANNING AND SCHEDULING OF A RESEARCH PROJECT Howard Eisner Repr. from Operations Res., v. 10, no. 1, Feb. 1962 p 115-125 refs

This paper describes a configurational and mathematical tool that may be used in the planning and scheduling of a research program. The configuration, termed a decision box (db) network, is a generalization of the PERT network and allows alternative procedures for accomplishing research tasks to be considered. The probability of following each alternative is estimated, leading to a ranking, in terms of probability of occurrence, of the possible outcomes of the research program. The entropy of the process is calculated using classical information theory concepts. Standard PERT techniques for time estimating and scheduling, superimposed upon the db network structure, complete the description of this research management tool. Author

N66-80359

Army Electronics Logistics Research Office, Phila

delphia, Pa. AN EXTENSION OF PERT TO MULTI-PROJECT AND FUNCTIONAL MANAGEMENT: PERT-C Jan. 1963 43 P Repr. of "A Signal Corps Concept for MultiProject Management" May 1962 (AD-417333)

This paper describes a management information system and evaluation technique which has been designed for use by the U. S. Army Electronics Command in planning, scheduling, and monitoring the acquisition of communications-electronic equipments. The technique is designed to be a response to the general problem of "project" or "systems" management across functionally oriented field operating agencies. One of its primary objectives is to provide data, in varying detail and format, for use at all levels of management-at local operating levels as well as at command levels. Author

N66-39602#. System Development Corp., Santa Monica, Calif. DEVELOPMENT OF EQUATION FOR ESTIMATING THE COSTS OF COMPUTER PROGRAM PRODUCTION

V. La Bolle Bedford, Mass., AFSC, Electron. Systems Div., Jun. 1966 57 p

(Contract AF 19(628)-5166) (TM-2918/000/00: ESD-TR-66-350; AD-637760) CFSTI: HC $3.00/MF $0.50

The report summarizes System Development Corporation (SDC) Technical Memorandum TM-2712 (AD-631 259). Additional sets of equations are given. Each set contains four equations; each equation shows how to form an estimate for one of the cost measures--number of man months, computer hours, new machine language instructions, months elapsed--by combining numerical values for selected factors that influence these costs. This report reviews the development of these equations including the application of statistical methods such as correlation and multivariate regression to experience data that characterize 74 computer programming efforts completed at SDC. The earlier work in the first cycle, a similar analysis of data for 27 SDC computer

programming efforts. is also described. After the publication of TM-2712, the second cycle was continued by additional analysis of the same SDC data for 74 computer programming efforts. The aim of the additional work was to improve the estimating precision of the equations presented in TM-2712 The improvements reported were achieved by deriving new cost equations, one set based upon a truncated sample and then three sets based upon three subsamples of the data. An interim evaluation of the work completed in the first and second cycles presents proposed improvements in approach and research methods Author (TAB)

for blending intuitive managerial judgments with the formal cost-estimation procedures. Supplementary findings include indications of the relative sensitivity of job cost to changes in the values for the indices, and preliminary comparisons of resource usage between programs produced in machine-oriented or procedure-oriented languages. Recommendations are made for the collection of more accurate and current data on programming jobs during the production cycle, and the development of a census of computer programming, to enable the design of precise sampling experiments for subsequent analyses. Author

N66-35641# Mitre Corp., Bedford, Mass.

AESOP: A GENERAL PURPOSE APPROACH TO REALTIME, DIRECT ACCESS MANAGEMENT INFORMATION SYSTEMS

Joseph Spiegel, John K. Summers, and Edward M. Bennett Bedford, Mass. AFSC, Electron, Systems Div.. Jun. 1966 38 p Presented at the Am. Management Assoc. Meeting. New York, Feb. 1966

(Contract AF 19(628)-5165)

(MTP-33; ESD-TR-66-289; AD-634371) CFSTI: HC $2.00/ MF $0.50

AESOP, a laboratory-based prototype of a general purpose. on-line, visually-oriented information system, is used to investigate ways of handling many different types and levels of command and management problems spanning organizational levels from the executive suite down through the staff and operations analysts to the actual system designers and programmers. In particular, it deals with those organizational activities that require highly flexible, direct-access capabilities, the system is configured for easy use by the inexperienced as well as by the sophisticated, and utilizes a variety of user station devices to facilitate such flexibility. including a cathode-ray-tube display, a lightgun, a typewriter, and associated push-buttons. At each station, it is capable of generating, editing, and formatting information on-line, as well as building. executing, and debugging on-line the analytic and mathematical procedures and algorithms of both the users and the system itself, depending upon the organizational area or level of the user. Although the basic prototype system was developed for use in military command and management planning and information systems, its philosophy and concepts are applicable to industrial and academic organizations.

Author (TAB)

N66-29484# System Development Corp., Santa Monica, Calif.

RESEARCH INTO THE MANAGEMENT OF COMPUTER PROGRAMMING: A TRANSITIONAL ANALYSIS OF COST ESTIMATION TECHNIQUES

G. F. Weinwurm and H. J. Zagorski Nov. 1965 221 prefs Includes errata sheet

(Contract AF 19(628)-5166)

(TM-2712/000/00: TM-2712/000/00A; AD-631259) CFSTI: HC $6.00/MF $1.25

The report embodies results of a continuing research effect on development of management guidelines, standards, and techniques of computer programming. The report focuses on a statistical analysis of 74 completed computer programming jobs in terms of their resource-costs and related variables. The primary results are: indices of job difficulty, job type, development environment, and job uniqueness; a costliness factor that permits programming tasks to be ranked in this respect; weighted composites of the indices for estimating the cost of particular programming jobs: and scoring and confidence-band techniques

N66-26411# Texas A&M Univ., College Station. Inst. of Sta

tistics. INTEGRAL OPERATORS FOR PERT SUB NETWORKS Technical Report No. 2

12 P ref

Larry J. Ringer 23 Mar. 1966 (Grant DA-31-124-ARO(D)-282) (AROD-4721-3; AD-630841) CFSTI: HC $1.60/MF $0.50 An algorithm was previously developed for computing the cumulative distribution of the completion time for certain PERT networks. These networks (which are termed 'multiple crossed') are composed of certain simple sub-networks: (a) two activities in series, (b) several activities arranged in parallel, and (c) five activities arranged in 'Wheatstone Bridge configuration. In this note an extention is made to the concept of a 'multiple crossed' network by adding two additional sub-networks as building blocks to the three mentioned. Included is the development of integral operators for the 'double Wheatstone Bridge and the 'criss-cross'. C.T.C.

N66-22554# Applied Physics Lab., Johns Hopkins Univ.. Silver Spring, Md.

SCHEDULE ANALYSIS PROCEDURE

R. P. Rich Apr. 1960 34 pref (Contract NOrd-7386)

(TG-397; AD-627398) CFSTI: HC $2.00/MF $0.50

The report describes a Schedule Analysis Procedure for use with the time schedule of a Research and Development (R and D) Program. The procedure is: (1) Identify certain clearly definable time points, called events, within the R and D program. (2) Identify the activities which must be carried out in order to get from each event to its immediate successors. (3) Obtain statistical estimates of the amount of time required for each of these activities. (4) Analyze the network of events and activities so defined to determine potential critical activities and chains of activities. (5) Take the necessary actionrescheduling, expediting, application of additional effort. etc. to relieve the criticality. (6) Repeat as the needs of the program indicate. Step 4 can be carried out on a digital computer in an expeditious fashion, but clearly the results of this step will be of little value unless the inputs derived from steps 1-3 are reasonably reliable, and the corrective action in step 5 is taken. Author (TAB)

[blocks in formation]

a cost analyst makes in estimating the costs of a military system. This report discusses five basic estimating methods, examines the advantages and disadvantages of each, and identifies the types of data required to use each method. The 5 methods are (1) per unit catalog price or planning factor, (2) cost-to-cost estimating relationship. (3) noncost-to-cost estimating relationship, (4) specific analogy, and (5) expert estimate.) Two basic data sources, documented historical evidence and projected expert opinion, are discussed as a means of carrying out these estimating methods. The report also briefly reviews existing data base systems that have been organized to make these data sources available to the cost analyst.

Author (TAB)

its economic possibilities; that is, at identical expenditures of labor, power, raw materials, semi-finished products, etc.. the same enterprise has the possibility of manufacturing different finished products. Each enterprise is put into relationship with a production map. The mathematical details and the concept of a differentiable map are given, and the conditions for the differentiability are formulated. The overall production possibilities of the entire economy are also described for a time period by the production map The criteria of optimality are considered for the aims of the economy, and formulas are derived to define management by means of evaluation, and the income and outlay structure. M.G.J.

N66-20637# Research Analysis Corp., McLean, Va. APPLICATION OF CHANCE-CONSTRAINED PROGRAMMING TO SOLUTION OF THE SO-CALLED "SAVINGS AND LOAN ASSOCIATION" TYPE OF PROBLEM

Ye. K. Degtyarev In its News of the Acad. of Sci. USSR. Dept. of Tech. Sci., Tech. Cybernetics, No. 6 11 Mar. 1966 p 96-112 refs (See N66-20606 10-10) CFSTI: $6.00

An algorithm for planning the fulfillment of a finite series of commands on asynchronous digital computer (ADC) is proposed if the commands consist of operations of known length which belong to a finite set of operations, and each operation may be fulfilled on a unique circuit from a finite set of circuits making up the ADC. As a result of application of the algorithm to a series of commands a linear graph is constructed with the vertex of the functions the critical path of which equals the time of fulfillment of the series of commands. A theorem is proved about the applicability of the algorithm. Author

N66-17591# David Taylor Model Basin, Washington, D. C.
Applied Mathematics Lab.

MODERN MISER: A CRITICAL PATH AND RESOURCE
ALLOCATION METHOD FOR THE UNIVAC LARC
Abel William Camara and Natalie Tarter Goldberg May 1964
142 p refs

(DTMB-1796, AD-602827) CFSTL: HC $5.00/MF $1.00

The MODERN MISER System is a planning and scheduling tool developed on the Remington-Rand UNIVAC LARC Computer. It is designed to assist management in comprehending the logical restrictions on a series of activities pertaining to one or many projects. The system includes a basic critical path method, a cost optimization routine, a float allocation procedure, the facility for manpower leveling, and the resource planning and scheduling method. Various algorithms used in the development of the system are discussed. The machine procedure used to adapt these formulas to the LARC are shown. Author (TAB)

N66-15196#

Joint Publications Research Service, Washing

ton, D. C MATHEMATICAL FORMULATION OF A SIMPLIFIED MODEL OF PRODUCTION PLANNING

Yu N. Tyurin 6 Jan. 1966 36 p refs Transl into ENGLISH from Ekon i Mat. Metody (Moscow), v. 1, no 3, May/Jun. 1965 p 391-409

(JPRS-33613, TT-65-30057) CFSTI $2.00

A production model is considered in which consumption during the planning period is given, and a production plan to satisfy this consumption is selected A system of objectively conditioned evaluations of commodities and production capacities was studied. Commodity is the term used to describe everything produced by the economic system, and the amounts of commodities are measured in continuously changing quantities. Each individual enterprise or economic unit is described by

N66-15167# Stanford Univ., Calif. Inst. for Mathematical Studies in the Social Sciences.

THE NONSUBSTITUTION AND NONSWITCHING THEOREMS IN A MODEL WITH FIXED CAPITAL

Eytan Sheshinski and Edwin Burmeister 20 Aug. 1965 28 p refs

(Contract Nonr-225(50)))

(TR-135: AD-622793) CFSTI: HC $2.00/MF $0.50

Let a competitive economy produce commodities of varying durabilities, such that (a) production processes exhibit constant returns to scale; (b) there is one exogenous nonproductible factor; (c) there are alternative techniques to produce each good; (d) it is possible to define conversion coefficients for old durable goods in terms of new goods of the same kind. Theorem: Let (A)-(d) hold. Then I: A long-run equilibrium of input-output coefficients and of prices in terms of wage units in uniquely determined for any preassigned value of the rate of interest. II: It is impossible to have identical techniques at different interest rates. This theorem generalizes Samuelson's static and dynamic nonsubstitution theorem. Author (TAB)

N66-11335# Pennsylvania Univ., Philadelphia.
INVENTORY CONTROL UNDER STOCHASTIC LEAD TIME
AND STOCHASTIC DEMAND

Kenneth R. Rand, Jr. (M.B.A. Thesis) 1965 111 p refs
(AD-620921)

An attempt is made to investigate the behavior of an inventory system in which lead time, the size of the demand order. and the time between successive demand orders are all random variables with known probability distributions. Since adequate analytical mathematical models are not existent, a computerbased simulation model is used to study the inventory system. An introduction to the inventory problem and a description of inventory systems currently in use are provided. The formulation of the model is described. Results are presented as graphs of stockout time as a function of reorder point. TAB

N66-10809#

Northwestern Technological Inst., Evanston, III. DATA, MODELING AND DECISION

A. Charnes and W. W. Cooper Jun. 1965 12 p refs Prepared jointly with Carnegie Inst. of Tech. Its Systems Res. Memo127

(Contracts Nonr-1228(10); Nonr-760(24)) (AD-620172)

Past experience with models-and related methods of analysis-may be an inadequate guide for managers considering a use of new tools now available. The latter, viewed as multiplevariable systems models, may differ in their data requirements and decision possibilities in comparison with predecessors that could handle only a few variables at a time. In approaching these new tools it is desirable to consider using the models as guides to data collection as well as decisions. This refers

not only to data variety but also to data quality as judged by reference to the model itself. It may then be possible to eliminate needless expenditures of time and money on collecting or refining data. It is also desirable to consider integrating the modelling and decision making. Evaluations may then be secured which can guide alterations to the model and also open new decision possibilities which would otherwise not be apparent. The value of such a joint approach to data, models and decisions is examined and illustrated in the following article with special reference to media mix and new products marketing applications. Author (TAB)

time management. These tools are initiated and reviewed by pure management organizations, and actually detailed and completed by the functional organizations. Integration tools of work management consist of event logic networks for work definition; program plans for agreement on how work will be done; systems requirements analysis for identifying and quantifying equipment, manning, and procedures; and interface control for responsibility assignment to insure compatible mating hardware. Integration tools of system schedule management consist of establishing schedules, keeping track of progress; and taking corrective action in system manageG.G.

ment.

N65-82741

System Development Corp., Santa Monica, Calif. MANAGEMENT OF COMPUTER PROGRAMMING FOR COMMAND AND CONTROL SYSTEMS

K. Heinze, N. Claussen, and V. La Bolle 8 May 1963 43 p refs (Contract ARPA SD-97)

(SDC-TM-903/000/02: AD-415721)

Programming managers, the Department of Defense, and the Air Force Systems Command are becoming increasingly concerned with improving programming and reducing costs. Thus, a study of management techniques for computer programming in command and control systems is being conducted at the System Development Corporation. In this document, the Computer Program Implementation Process (CPIP) project reports the findings of this study. Seven computer programming efforts were surveyed to reveal the characteristics of computer programming, including growth, military use, and staffing. In a section titled "Program Development Problems," such management difficulties as computer unavilability and the effects of customer delay in design decisions are discussed. The CPIP project recommends improving: (1) the identification of programming activities and products by, among other means, establishing a common technical language and, (2) customer relations, by recognition and understanding of customer needs (including his working environment), and the need for the customer to understand the developer's approach to the program. Managers have difficulty in controlling and planning programming efforts without precise and detailed cost data, standard performance measures, and definitions of tasks and products. Knowledge of managing and developing computer programming systems must be extended and detailed, and programming must be formalized. Author

[blocks in formation]
[blocks in formation]

T. A. J. Nicholson Mar. 1965 30 p refs (AERE-M-1568) HMSO: 4s

A procedure is described to schedule jobs through a workshop. Each of the jobs consists of a specified sequence of operations, and the procedure is designed to order these operations through the machines efficiently in terms of a management objective. The usual objective is to complete the work in the minimum time, but in this paper three objectives are distinguished. The procedure does not necessarily provide an optimal scheduling, but it should be near optimal. It is essentially simple and easy to implement. Author

N65-23971# Boeing Co., Seattle, Wash.

THE INTEGRATION TOOLS OF SYSTEM MANAGEMENT H. K. Hebeler Paris, Eurospace. [1964] 46 p Presented at the U.S.-European Conf., Rome, 22-24 Jun. 1964

Discussed are the major tools of system management which relate to the integration of the work of one contractor with the work of another, with emphasis in the areas of work and

N65-23091# Tufts Univ., Medford, Mass.

COMPUTER SIMULATION MODEL FOR ORGANIZATION THEORY

Thornton B. Roby [1964] 78 p refs

(Contracts Nonr-494(15); AF 19(628)-2450) (TR-8: AD-611870)

A theoretical discussion is presented on the utilization of computers in analyzing organization theory. The computer simulation model is considered a novel and unique theoretical tool. Three aspects of the model are considered: the structure of the initial nucleus from which the model is refined and elaborated; the substantive elements or building blocks of which the model is constructed; and the language in which the model is couched. The process of testing and revising the computer model, and its relationship to empirical data, are discussed. A four-stage testing process is outlined; internal testing, initial tests of ad hoc validity, tests of extended validity or generalizability, and process evaluation. TAB

N64-84874 Army Signal Corps., Philadelphia, Pa. Logistics Evaluation Committee.

A SIGNAL CORPS CONCEPT FOR MULTI-PROJECT MANAGEMENT

William K. Kreamer May 1962 54 p

(Proj. Comet)

(AD-276938)

This paper describes a management information system and evaluation technique which has been designed for use by the U. S. Army Signal Corps in planning, scheduling, and monitoring the acquisition of communications-electronic equipments. The technique is designed to be a response to the general problem of "project" or "systems" management across functionally oriented field operating agencies. One of its primary objectives is to provide data, in varying detail and format, for use at all levels of management at local operating levels as well as at command levels. Author

[blocks in formation]

Three broad aspects of information system simulation and modelling are discussed: a methodological spectrum that has proved to be useful in understanding the role of various techniques in systems research; experimental design in empirical systems research, illustrated by a recently completed simulation study; functional task analysis of the information research project at the "microscopic" level to facilitate the growth of an information systems science. Also discussed is the Leviathan project which is a unique, experimental approach to studying the structure and social dynamics of large-scale organizations such as a military command, a government bureau, or an industrial organization. Three simulations are discussed: a automatic-mode simulation, a live simulation, and a dual-mode simulation.

Author

N64-12319 Raytheon Co., Waltham, Mass.
PERTeach. THE DEVELOPMENT OF EXPERIMENTAL PRO-
GRAMS FOR AUTOMATED TRAINING IN DECISION MAK-
ING

G. J. Rath, R. E. Hannan, and M. L. Weiss L. G. Hanscom Fid.,
Bedford, Mass., Decision Sci. Lab., Aug. 1963 40 p refs
(Contract AF 19(628)-365)

(ESD-TDR-63-608; AD-421735)

This report discusses the concepts and techniques employed in the preparation of the programed instruction courses for training management and support personnel in PERT. The applicability of PERT to decision making is also discussed.

Author

[blocks in formation]

A BRIEF REVIEW OF SIMSCRIPT AS A SIMULATING TECH-
NIQUE

M. A. Geisler and H. M. Markowitz Aug. 1963 27 p 3 refs
(Contract AF 49(638)-700; Proj. RAND)
(RM-3778-PRI)

The underlying concepts, structure, and current status of SIMSCRIPT are reviewed. SIMSCRIPT is a way of designing and writing down a simulation model, and has been developed with the following objectives: (1) to produce a generalized structure for designing simulation models; (2) to provide a rapid way of converting a simulation model into a computer program; (3) to provide a rapid way of making changes in the simulation model which can be readily reflected in the machine program; and (4) to provide a flexible way of obtaining useful outputs for analysis. The method is not only an abridged language, but also a structure with the help of which a wide class of management problems can be programed to a computer. It is SO designed that whole areas of a problem can be changed without reprograming the entire model. SIMSCRIPT streamlines programing and makes it faster and more economical to use. N.E.A

N63-15702 RAND Corp., Santa Monica, Calif.
USES OF MONTE CARLO IN PERT
Richard Van Slyke Feb. 1963 28 p 9 refs
(Contract AF 49(638)-700; Proj. RAND)
(RM-3367-PR)

The mathematical assumptions underlying the PERT (Program Evaluation and Review Technique) management system are of doubtful validity, and, even granting these assumptions, there are still serious computational difficulties involved in getting the desired answers. This report outlines some of the weaknesses of the present system, and suggests how the use of Monte Carlo methods can lead to improvement, first in allowing less restrictive mathematical assumptions to be made, and second in extending the kinds of computational results that can be obtained. Moreover, the method can be used to check the validity of the commonly used approximations. A "criticality index" for an activity is defined: it is the probability that an activity will lie on a critical path. This is an example of a quantity that can be calculated by Monte Carlo methods, but not by presently used methods. An experimental computer program for Monte Carlo treatment of PERT networks has been coded for the IBM-7090 computer. The program is discussed, and some shortcuts are suggested for reducing the time of computation. Author

N63-13146 Matrix Corp., Arlington, Va.

DEVELOPMENT OF QUALITATIVE AND QUANTITATIVE PERSONNEL REQUIREMENTS INFORMATION [Final Report] Walter L. Smith, Melvin T. Snyder (Behavioral Sciences Lab.) Robert G. Demaree, and Melvin R. Marks Wright-Patterson AFB, Ohio, Aerospace Med. Res. Labs. (6570th), Dec. 1962 88 p 211 refs (Contract AF 33(616)-7464) (MRL-TDR-62-4) OTS: $2.25

This report describes the nature of, purpose, and method for developing Qualitative and Quantitative Personnel Requirements Information (QQPRI). The nature and scope of QQPRI and the relationship of QQPRI to system development are covered. The general constraints within which QQPRI is developed are described. Such constraints include technical weapon system constraints, as well as policies on Air Force Personnel, Training, and Manning. Emphasis is placed on QQPRI development techniques such as task analyses, estimation of performance time, establishing Air Force positions, and determining the relation of these to existing Air Force specialties. Author

« ÀÌÀü°è¼Ó »