Main Menu



Block Schedule


Meeting Committee


Industrial Statistics / Statistique industrielle
(N. Balakrishnan, Organizer)

MANOU HOSSEINI, Global Management Science Services, Toronto, Ontario  M3A 3R8
Performance-dependability prediction in an uncertain environment

Maintenance optimization and reliability management of technical systems is an important area of the application of industrial statistics. Due to various factors such as variability in: material, manufacturing, installation, operators and maintainers' skills, no two nominally identical equipment show exactly the same survival or time to failure. Moreover, in a global analysis or performance-dependability evaluation of a manufacturing system there are other sources of uncertainty that should be dealt with statistical and probabilistic approaches. Variables such as duration of inspection, repair, and maintenance activities, processing time, switch from a standard state to a substandard state, availability of resources, and market demand are not deterministic. In this paper a hybrid stochastic model for development of maintenance strategies, that assure both best performance and dependability of an operating system, is demonstrated.

Controller performance monitoring in the steel industry

A typical steel manufacturing facility contains thousands of feedback control loops. When a feedback control system is well designed and tuned, the manipulated control elements adequately compensate for the effects of disturbances in order to reduce the costs of poor product quality. In many situations, however, the application of feedback controllers can contribute negatively to the level of process variability. This can stem from a controller that is designed approximately correctly but that is miss-tuned for example. When a standard three sigma Shewhart chart on a process quality measurement alarms, an additional controller monitoring chart can be checked to ascertain if the associated controller is operating well, and an investigation can be made to see if any modifications can be made to the control scheme to solve the apparent quality problem. Feedback controllers can be monitored via a scheme based on the control performance index (CPI) and its associated variance. The CPI index computes the ratio of the present process variability to the variability that could be attained with a useful theoretical benchmark called minimum variance control. This monitoring method is presented, and an example of a preliminary application on a process in an integrated steel mill is discussed.

MIKE G. RICHARDS, Kodak Canada Inc., Toronto, Ontario
Multivariate process analysis-advanced tool for six sigma quality

Six Sigma is a measure of the capability of a process to deliver defect free product. A process operating at a six sigma level of process capability produces only 3.4 defects per million opportunities.

This presentation provides an overview of Six Sigma concepts and the role of the Six Sigma Black Belt practitioner in leading high impact quality improvement initiatives.

Multivariate process analysis is introduced as an advances Black Belt tool that can be used to develop new process understanding needed to guide breakthroughs in process and quality performance.

SAMUEL SHEN, Department of Mathematical Sciences, University of Alberta, Edmonton, Alberta  T6G 2G1
Optimal estimation of climate parameters

This talk discusses two optimal methods for assessing a climate state. The first is the spectral approach to optimal averaging of the historical climate data. An optimal averaging scheme minimizes the mean square error and can be used to measure various orders of spherical harmonic components of a climate field with finitely many surface stations. An important formula was derived to demonstrate that the sampling error is relatively insensitive to the exact shapes of empirical orthogonal functions. Two examples are described: the global average of the annual surface air temperature using 63 stations and the regional average of the monthly tropical Pacific sea surface temperature. The second is the adaptive gridding method for the historical climate data. Validation of climate models requires the reconstruction of climate fields of the past, say 1885-1930, from the scarce observed data. A field can be reconstructed on a one-degree lat-long grid. A systematic theory for the interpolation is described and it uses the emperical orthogonal functions computed from the recent and more accurate observed data. The last part of the talk will be on the opportunities of mathematics/statistics/computing in climatic, agricultural and environmental research.


    1. S. S. Shen, G. R. North and K. Y. Kim, Spectral approach to optimal estimation of the global average temperature. J. Climate 7(1994), 1999-2007.
    2. T. M. Smith, R. E. Livezey and S. S. Shen, An improved method for interpolating sparse and irregularly distributed data onto a regular grid. J. Climate 11(1998), 1717-1729.


top of page
Copyright © 2000 Canadian Mathematical Society - Société mathématique du Canada.
Any comments or suggestions should be sent to - Commentaires ou suggestions envoyé à: