The interface of computer science and statistics

Current scientific, technical, and management progress is characterized by the generation of a tremendous amount of data for analysis. This, in turn, poses a significant challenge: to effectively and efficiently extract meaningful information from the large volume of data. Two relatively young professions, computer science and statistics, are intimately linked in any response to the challenge. They have consequently become indispensable to scientific, technical, and management progress, occupying a position at its very heart Computer science and statistics have each been separately documented by many books as well as numerous papers. However, the interface of computer science and statistics, the area of their interaction, has been documented only in part. This paper begins characterization of the entire interface by providing a structure and an historical background for it A structure for the interface is introduced initially, followed by an historical background for the interface presented in two parts. First to be summarized is the evolution of the interface from an interweaving of the mechanical prerequisites to the computer and mathematical prerequisities to computer science and of the foundations for probability and statistics. Development of statistics prior to 1900 then is reviewed.     

Optimal policy for a dynamic multi-echelon inventory model

A general multiperiod multi-echelon supply system consisting of n facilities each stocking a single product is studied. At the beginning of a period each facility may order stock from an exogenous source with no delivery lag and proportional ordering costs. During the period the (random) demands at the facilities are satisfied according to a given supply policy that determines to what extent stock may be redistributed from facilities with excess stock to those experiencing shortages. There are storage, shortage, and transportation costs. An ordering policy that minimizes expected costs is sought. If the initial stock is sufficiently small and certain other conditions are fulfilled, it is optimal to order up to a certain base stock level at each facility. The special supply policy in which each facility except facility 1 passes its shortages on to a given lower numbered facility called its direct supplier is examined in some detail. Bounds on the base stock levels are obtained. It is also shown that if the demand distribution at facility j is stochastically smaller (“spread” less) than that at another facility k having the same direct supplier and if certain other conditions are fulfilled, then the optimal base stock level (“virtual” stock out probability) at j is less than (greater than) or equal to that at facility k.     

Discounted production scheduling and employment smoothing

We consider the problem of minimizing the sum of production, employment smoothing, and inventory costs over a finite number of time periods where demands are known. The fundamental difference between our model and that treated in [1] is that here we permit the smoothing cost to be nonstationary, thereby admitting a model with discounting. We show that the values of the instrumental variables are nondecreasing in time when demands are nondecreasing. We also derive some asymptotic properties of optimal policies.     

A note on a computational model for a data/voice communication queueing system

Certain types of communication nodes can be viewed as multichannel queueing systems with two types of arrival streams. Data arrivals are characterized by high arrival and service rates and have the ability to queue if all service channels are busy. Voice arrivals have small arrival and service rates and do not have the ability to wait when the channels are full. Computational procedures are presented for obtaining the invariant probabilities associated with the queueing model.     

A complete importance ranking for components of binary coherent systems,with extensions to multi-state systems

Means of measuring and ranking a system's components relative to their importance to the system reliability have been developed by a number of authors. This paper investigates a new ranking that is based upon minimal cuts and compares it with existing definitions. The new ranking is shown to be easily calculated from readily obtainable information and to be most useful for systems composed of highly reliable components. The paper also discusses extensions of importance measures and rankings to systems in which both the system and its components may be in any of a finite number of states. Many of the results about importance measures and rankings for binary systems are shown to extend to the more sophisticated multi-state systems. Also, the multi-state importance measures and rankings are shown to be decomposable into a number of sub-measures and rankings.     

Coordinated replenishments of items under time-varying demand: Dynamic programming formulation

We consider a group (or family) of items having deterministic, but time-varying, demand patterns. The group is defined by a setup-cost structure that makes coordination attractive (a major setup cost for each group replenishment regardless of how many of the items are involved). The problem is to determine the timing and sizes of the replenishments of all of the items so as to satisfy the demand out to a given horizon in a cost-minimizing fashion. A dynamic programming formulation is illustrated for the case of a two-item family. It is demonstrated that the dynamic programming approach is computationally reasonable, in an operational sense, only for small family sizes. For large families heuristic solution methods appear necessary.     

A partitioning technique for obtaining solutions to the modularization problem

A significant problem in electronic system design is that of partitioning the functional elements of an equipment schematic into subsets which may be regarded as modules. The collection of all such subsets generated by a particular partitioning forms a potential modular design. The specific problem is to determine that partitioning of the schematic that minimizes a cost function defined on the subsets subject to specified hardware, design, packaging, and inventory constraints. This problem is termed the modularization problem. This paper presents a method for obtaining restricted solutions to the modularization problem by employing some recent developments in linear graph theory obtained by one of the coauthors. Numerical results from the solution of several typical problems are presented.     

Optimization in mixed-integer space with a single linear bound

The search for an optimal point in a mixed-integer space with a single linear bound may be significantly reduced by a procedure resembling the Lagrangian technique. This procedure uses the coefficients of the linear bound to generate a set of necessary conditions that may eliminate most of the space from further consideration. Enumerative or other techniques can then locate the optimum with greater efficiency. Several methods are presented for applying this theory to separable and quadratic objectives. In the maximization of a separable concave function, the resulting average range of the variables is approximately equal to the maximum (integer) coefficient of the constraint equation.     

Maintenance policies when failure distribution of equipment is only partially known

An optimal schedule for checking an equipment subject to failure which can be detected by inspection only, is derived. Increasing failure rate and one percentile specify the otherwise unknown life distribution. Dynamic programming methodology yields the solution which minimizes the maximum expected cost. Numerical examples are presented and compared with models employing differing amounts of knowledge.