M. I. Dessouky

Strategic Behavior, Workload, and Performance in Task Scheduling

Scheduling theory is proposed as a normative model for strategic behavior when operators are confronted by several tasks, all of which should be completed within a fixed time span, and when they are free to choose the order in which the tasks should be done. Three experiments are described to investigate the effect of knowing the correct scheduling rule on the efficiency of performance, subjective workload, and choice of strategy under different conditions of time pressure. The most potent effects are from time pressure. The reasons for the weak effect of knowing the rules are discussed, and implications for strategic behavior, displays, and decision aids are indicated.

The One-Machine Sequencing Problem with Early Starts and Due Dates

Abstract An optimal procedure is developed for the sequencing of n-jobs on one machine to minimize the maximum job lateness when the n-jobs have unequal early start times and due dates. The decision rule derived for the procedure primarily sequences the jobs in order of their nondecreasing early starts plus due dates. A branch-and-bound approach is employed by the procedure which moves to opti-mality in a finite number of steps.

An application of multiple criteria decision making principles for planning machining operations

Abstract Some recently introduced concepts of multiple criteria decision making (MCDM), such as the range of compromise and its use in dimensionless weighting of the objectives, are employed to formulate machining recommendations under multiple machining criteria. Two approaches are presented and applied to the machining of 390 die cast aluminum alloy. The first is a graphical presentation of nondominated solutions from which an appropriate solution can be chosen under varying applications, requirements, and priorities. The second is the formulation of a composite function of the objectives through dimensionless weighting to select the optimum machining variables under given aspiration levels for the machining criteria.

A Forward-Backward Procedure for the Single Machine Problem to Minimize Maximum Lateness

Abstract This paper presents an optimal procedure for sequencing n jobs on one machine to minimize the maximum lateness when they may have different ready times, processing times, and due dates. This procedure is significantly more efficient for difficult problems than the McMahon/Florian algorithm which has achieved the best results in this area to date. The primary measures of computational efficiency for these procedures are the mean computer processing time and the percentage of problems solved optimally within a given time limit. An experiment that employs response surface methodology to locate problems within the region of highest difficulty shows the comparative efficiency of this procedure.

Heuristic procedures for the single machine problem to minimize maximum lateness

Abstract Eleven heuristic procedures are presented for sequencing n jobs on one machine to minimize the maximum job lateness when the jobs may have different ready times, processing times, and due dates. The performances of seven procedures that are shown to be representative of all eleven are evaluated on sets of sample problems against optimal solutions derived by a branch-and-bound algorithm. The main measures of each procedures comparative quality are (1) the mean number of time units by which its solution exceeds the optimal solution and (2) the percentage of times that an optimal sequence is attained. The motivation for employing heuristic procedures is to obtain a sequence that will provide a satisfactory practical response to real world scheduling problems with minimum computational effort.

An integrated quality systems approach to quality and productivity improvement in continuous manufacturing processes

An integrated quality systems methodology is presented as a framework within which the concepts of process control can be used to improve quality and productivity. The process is mathematically described by stochastic time series models which statistically describe how inputs and outputs interact. Several different methods for fault identification, including autocorrelation checks of the model residuals, forecasting prediction intervals, and the cusum chart are compared in terms of relative performance. A helix cable manufacturing process is simulated and analyzed by the methodology and faults are identified and suggestions are made for process improvement. Through the simulation these time series control chart methods are shown to be much more effective than conventional methods such as Shewhart control charts.

Symmetry and optimally properties of the single machine problem

Abstract This paper develops some useful concepts and properties for the problem of sequencing n jobs on one machine so as to minimize the maximum lateness when the jobs may have unequal ready times, processing times, and due dates, and pre-emption is not permitted. The properties established include the equivalence among three versions of the problem, the symmetrical nature of the problem and the additive and multiplicative characteristics of the job set parameters. These properties are applied to forward and backward sequencing. A strong sufficient optimality condition is developed and used to obtain tight lower bounds on the optimal solution. The paper demonstrates the applicability of the concepts and properties to the development of efficient optimal and heuristic procedures and to the design of experiments for evaluating the relative performance of solution procedures. Optimal single-pass solutions for special cases of the problem are given.

Cost-Duration Analysis with the Cut Network

Abstract A new approach is developed for conducting a cost-duration analysis (trade-off study) for a project denned with a CPM/PERT activity network. The approach utilizes a new graph theory concept, proper oriented cut-set (POCS). It also uses the cut network, a dual-type network defined in this paper for any planar activity network. The cut network implicitly enumerates all POCSs of an original network. The cost-duration problem defined for the original activity network is transformed into a minimum-cost flow problem for the cut network. This approach to cost-duration analysis, developed here for a planar network and linear activity cost-duration relations, is optimal, conceptually direct, and has potential for application to more general cost-functions.