Research for Computational Methods

Syndicate content
Fixed Shortage Costs and the Classical Inventory Model
June 1, 1974
Many recent economic and inventory studies have included various types of fixed or lump-sum costs as important determinants of optimal behavior. In this paper, the classical inventory model is augmented to include fixed shortage costs. In general, the presence of fixed shortage costs can lead to complex optimal solutions. The purpose of this paper is to establish a set of sufficient conditions which guarantee the existence of an optimal ordering policy which is unique. The resulting optimal policy is described by a unique set of critical numbers which are bounded and decrease monotonically over the horizon for which the inventory system is to be operated.
Read More | Download Report
FISHER EXACT PROBABILITY TEST
June 1, 1974
The Fisher Exact Probability Test is a non-parametric technique for testing whether or not two samples could reasonably be expected to have originated from the same population. This paper defines the test, demonstrates its use, and explains a Fortran program which was developed to calculate the required values.
Read More | Download Report
Inventory and the Theory of the Firm
June 1, 1974
A serious criticism of the inventory models presented in the literature is that they often seek to minimize inventory costs rather than to minimize total costs for a specified production level. This neglect is caused, in large measure, by a failure to consider inventories as factors of production, with substitution possibilities between other types of inputs. This paper incorporates these substitution possibilities into the analysis, and formulates an inventory model within the theory of the firm.
Read More | Download Report
ESTIMATION OF AIM-ERROR CORRELATION, AIMING DISPERSION AND BALLISTIC DISPERSION
June 1, 1974
Assuming the the x-coordinates of impact points of rounds fired at a target centered at the origin, x=0, are the sums of correlated aim errors and independent ballistic errors, this report shows how the likelihood function of the unknown statistical parameters, associated with a sample x1,x2 ..., xN of impact points from a single burst, can be quickly and simply computed. The likelihood function is used here as the basis for estimating the parameters, and the results of the estimates, applied to a large number of computer simulations, are summarized.
Read More | Download Report
COMPUTER PROGRAMS FOR ANALYZING MANPOWER DATA
June 1, 1974
The MANPOWER program file consists of computer programs developed to facilitate machine extraction and processing of Five-Year Defense Program and other manpower data. The file includes routines for data file building, manipulation, and data transformation and analysis. A set of utility routines is described which allows the user to extract and execute any program on the MANPOWER file without having to handle any of the source program card-decks. Using these routines, the MANPOWER program file, and a data file, the entire process of building and analyzing a data bank could be accomplished as a single job. This system has been developed for use on Control Data 3400, 3600, and 3800 machines. Modificatiaons to some of the programs might be necessary if other equipment were used.
Read More | Download Report
RELATING FACTOR ANALYTICALLY DERIVED MEASURES TO EXOGENOUS VARIABLES
June 1, 1974
This paper presents a method by which factors and exogenous variables can be correlated without computing factor scores.
Read More | Download Report
SENSITIVITY TESTS ON SLAT COMPUTER SIMULATIONS USING EXPERIMENTAL DESIGN
June 1, 1974
Experimental designs are proven to be just as useful in studying the interrelations between variables in computer simulations as they have been in agriculture and industry. This paper demonstrates that the statistical experimental design technique aids the analysis of complex simulation by efficient selection of the computer runs to be made and thereby saves computer time. Further the analysis is more complete, showing that the variables act jointly, and measures the goodness of fit of predicted values. As an example, the sensitivity tests on changes in air sortie effectiveness on a land campaign are discussed. The techniques, terminology and calculation methods with a bibliography are discussed in an appendix for those who wish to learn more about the technique.
Read More | Download Report
PROBABILITIES IN A SEQUENCE OF CORRELATED EVENTS
June 1, 1974
The correlation between gunnery shots is treated by regarding them as repeated dependent trials, the probability of a hit being determined by the success or failure of the previous shot. Relevant results on MARKOV chains are summarized and extended to the case of non-stationary correlation. The system of difference equations is made the basis of the treatment, which provides an IBM computational scheme for finding the probabilities of various numbers of hits, and also explicit formulas for such probabilities in a case which generalizes the Poisson distribution. Illustrative applications are given to gunnery data.
Read More | Download Report
ON BEHAVIOR STRATEGY SOLUTIONS IN TWO-PERSON ZERO-SUM FINITE EXTENDED GAMES WITH IMPERFECT INFORMATION, PART I: A METHOD FOR DETERMINATION OF MINIMALLY COMPLEX BEHAVIOR STRATEGY SOLUTIONS
June 1, 1974
This paper provides an analysis of the relationship between any given two-person zero-sum finite extended game with given information schemes for its players, and an associated N-person non-cooperative game.
Read More | Download Report
LINKING A VARIABLE WEIGHT SERIES
June 1, 1974
It has long been accepted that in many countries - especially those in a developing stage - frequent revisions of the weighting pattern of the Consumer Price Index are called for. It is further argued that it is impractical to change the base every time the weights are revised and so the new series should be linked into the existing one. This paper discusses the problems caused by using any particular month to link a series which has variable weights. It suggests that a series of this type requires a separate linking constant for each set of weights.
Read More | Download Report