doi:10.1016/j.ejor.2005.05.010 
Copyright © 2005 Elsevier B.V. All rights reserved.
Discrete Optimization
Aversion scheduling in the presence of risky jobs
Gary W. Blacka, 
, 
, Kenneth N. McKayb and Thomas E. Mortonc
aSchool of Business, University of Southern Indiana, 8600 University Boulevard, Evansville, IN 47712, USA
bDepartment of Management Sciences, University of Waterloo, Waterloo, ont, N2L 3G1, Canada
cGraduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, PA 15213, USA
Received 22 September 2003;
accepted 13 May 2005.
Available online 10 August 2005.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
Empirical studies have shown that human schedulers use special procedures to deal with troublesome jobs that are perceived to disrupt manufacturing or that will take substantially longer than the industrial engineering standards. These troublesome jobs present a risk to the manufacturing process and to the schedule robustness. One strategy is to delay them whenever possible and to allow other work to overtake. The Aversion Dynamics concept is used to include this type of logic in scheduling heuristics such that a trade-off analysis of penalties occurs in light of the expected performance results. This is accomplished by altering processing time estimates to achieve a form of “safety time” for risky jobs. This paper conducts a large empirical study within the single-machine static arrival environment to demonstrate that the concept of special sequencing based on job risk is significant and that robust strategies can be developed.
Keywords: Scheduling; Aversion Dynamics; Adaptive heuristics; Risk mitigation heuristics
Fig. 1. Best safety time versus log10(weighted tardiness)—mean risk level known.
Fig. 2. % Penalty using ST = 0 versus log10(weighted tardiness), low risk, mean known.
Fig. 3. % Penalty using ST = 0 versus log10(weighted tardiness), medium risk, mean known.
Fig. 4. % Penalty using ST = 0 versus log10(weighted tardiness), high risk, mean known.
Table 1.
Summary of weighted flow (lateness) results (negative exponential risk) (similar to weighted tardiness results whenever TF = 1.0)
Table 2.
Summary of weighted tardiness results for TF = 0.7 (negative exponential risk) (averaged across 4 due date range (RDD) settings)
Table 3.
Summary of weighted tardiness results for TF = 0.5 (negative exponential risk) (averaged across 4 due date range (RDD) settings)
Table 1.1
Low risk, mean known
Table 1.2
Medium risk, mean known
Table 1.3
High risk, mean known
Table 1.4
Low risk, mean unknown (3 observations)
Table 1.5
Medium risk, mean unknown (3 observations)
Table 1.6
High risk, mean unknown (3 observations)
Table 1.7
Low risk, mean unknown (1 observation)
Table 1.8
Medium risk, mean unknown (1 observation)
Table 1.9
High risk, mean unknown (1 observation)
Table 2.1
Summary of weighted flow (lateness) results (uniform risk) (equivalent to weighted tardiness results when TF = 1.0)
Table 2.2
Summary of weighted tardiness results for TF = 0.1, RDD = 0.3 (uniform risk)
Table 2.3
Summary of weighted tardiness results for TF = 0.3, RDD = 0.3 (uniform risk)
Table 2.4
Summary of weighted tardiness results for TF = 0.5, RDD = 0.3 (uniform risk)
Table 2.5
Summary of weighted tardiness results for TF = 0.7, RDD = 0.3 (uniform risk)
Abstract
Purchase PDF (239 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (0)
Purchase PDF (159 K)
