Contents lists available at ScienceDirect Computers & Industrial Engineering journal homepage: www.elsevier.com/locate/caie Single-measure and multi-measure approach of predictive manufacturing control: A comparative study Minakshi Kumari a, ⁎ , Makarand S. Kulkarni b a Department of Mechanical Engineering, Indian Institute of Technology Delhi, India b Department of Mechanical Engineering, Indian Institute of Technology Bombay, India ARTICLEINFO Keywords: Complexity CRITIC (Criteria Importance through Intercriteria Correlation) MCDM (Multi criteria decision making) Proactive control Simulation Uncertainty modeling ABSTRACT The paper presents a comparative study of two predictive control approaches in the context of a manufacturing shop foor. The underlying schema of the current work is to develop a control mechanism for shop foor that is able to predict the expected behaviour of environmental variables on a projected timeline and hence identify those which need intervention. In order to execute this, a multi-measure approach and a single-measure ap- proach have been proposed. The multi-measure approach comprises of complexity, penalty and desirability as the three criteria and the single-measure approach deploys a unifed index based mechanism to aide decision making. The single -measure approach uses CRITIC (Criteria Importance through Intercriteria Correlation); a multi criteria decision making method, for the unifcation of criteria. The reported study compares the efciency of the two approaches in signifcant variable identifcation in similar shop foor scenarios for a multi-component single machine system. The work is a simulation based study with a case developed out of a real life high pressure die casting unit. The results demonstrate that the unifcation of criteria causes a dampening efect on the sensitivity of the control measure. As a result, the number and nature of environmental variable identifed as signifcant is low in single-measure approach. The reported comparative study will facilitate an informed choice between the two approaches depending on the extent of control a system seeks. 1. Introduction Importance of predictive control as an enabler to augment perfor- mance is time and again agreed upon by manufacturing organisations. Unlike the traditional approach of reactive control, which attempts to bring things into order once an undesirable event has triggered, the predictive approach aims at beforehand assessment of all possible “what if” scenarios and propose a “set of action items”. The latter approach is an active means of averting any detrimental situation to the best pos- sible extent. In the context of a typical manufacturing shop foor, pre- dictive control is not only complex but has a deep and an elaborate sense too. It is complex because it encompasses various shop foor functions i.e. production scheduling, maintenance and quality with each of them extending an inter-functional infuence.Anditisdeepand elaborate as it can actively infuence the business strategy and dictate an organisation’s market presence. Having discussed the importance of predictive control of shop foor, the question at hand is, “what measure (s) needs to be predicted to facilitate efective control”? Also, “does it need to be a single or a multi-measure approach of control”? The current work presents a comparative study of a single-measure versus a multi-measure approach of proactive shop foor control. The work reported is a sequential set of studies that started with modeling of control measures with the perspective of a shop foor manager which then extended to an analysis of inter relation of the proposed measures and their subsequent unifcation. The results reported present a com- parison of using a multi-measure against a unifed control measure approach for a single machine system. The study has been conducted in a high pressure die casting manufacturing unit with three machines of same model with diferent component age. A periodic simulation ap- proach has been deployed for analysis. Thoroughness with which the models (both multi -measure as well as unifed index) have been built and their readiness for use on any shop foor are the highlights of the current work. The comparative study is an attempt to bring out the insights as to how and when to use one type of index over the other in diferent manufacturing set ups and scenarios. The paper is organized in sections with Section 2 discussing the problem statement. Section3 presents the literature reviewed. Section4 givesanoutlookofthereferencesystemforwhichthepresentstudyhas https://doi.org/10.1016/j.cie.2018.12.018 Received 16 March 2018; Received in revised form 8 November 2018; Accepted 5 December 2018 ⁎ Corresponding author at: Industrial Engineering Lab, Block II-432, Department of Mechanical Engineering, Indian Institute of Technology-Delhi, Hauz Khas, New Delhi 110016, India. E-mail address: minakshik@mech.iitd.ac.in (M. Kumari). Computers & Industrial Engineering 127 (2019) 182–195 Available online 07 December 2018 0360-8352/ © 2018 Elsevier Ltd. All rights reserved. T