A Hybrid AI-Based Conceptual Decision-Making Model for Sustainable Maintenance Strategy Selection

Introduction

Classical theory of maintenance engineering has been based on the paradigm of fixing broken down assets to bring them back to the operative condition(s). However, even since the advent and advantages of several other improved paradigms of maintenance engineering, this ancient philosophy has been believed to be beneficial for non-critical or single shot items or where the repair is not an economically viable option. Therefore, for today’s sophisticated and critical systems/machinery such as aerospace engines, boilers, high speed gearboxes, the earlier concept of maintenance engineering has transitioned towards a more advanced philosophy, where it is not only considered from a technical perspective (i.e., vibration monitoring, temperature monitoring, visual inspections, etc.), but is extended towards managerial activities (i.e., planning, coordination, personnel management, etc.) (Shafiee, 2015). Therefore, in the broader sense, it can be redefined as a set of activities comprising of technical, administrative, managerial routines during the life cycle of an asset and is carried out to preserve their values in terms of reliability, maintainability, availability and safety, which in turn directly influence the productivity and reputation of an organization. Adding to this context, it has been reported that depending on the type and size of organizations, a substantial percentage of the annual budget, varying from 15% to 60%, goes to deliver the maintenance program of an organization (Nezami & Yildirim, 2013; Mobley, 2002).

Now-a-days, most of the organizations demand that the working critical systems/machinery must have high availability and reliability, failing which might lead to a substantial amount of losses in terms of outgoing products’ quality, generated revenue, missing targeted output, and even human casualties and environmental damages followed by litigations and lawsuits for the occurrence of certain types of failures. To avert such unwarranted failures of these systems and consequences thereof, they must be abetted with optimal maintenance strategies which not only reduces their annual maintenance cost but also alleviates them from unwanted operational inefficiencies and poor performance.

However, choosing the optimal maintenance strategy for critical systems/machinery is considered a complex decision-making task, where, at first, organizations have to identify a group of potential experts who have the required level of expertise and domain knowledge. Thereafter, a set of technical and managerial factors (e.g., criticality of operation of the system, working condition, level of operators' involvement, environmental-social-economic impacts, etc.) are carefully defined by those expert panelists and in the very next step, based on those chosen factors, each of the maintenance strategies (e.g., failure based maintenance, time-based maintenance, reliability centered maintenance, condition based maintenance, total productive maintenance, etc.) are rated accordingly, either in a subjective or objective manner. Upon completion of these, mathematical decision-making tool(s) are employed to arrive at the decision. It is worth to mention that the chosen factors are always conflicting in nature (i.e., if the value of a factor upsurges then another one may rise or decline), which makes the situation more complex to arrive at the optimal decision.

The Maintenance Strategy Selection Problem (MSSP) can be stated, just like any other optimization problem with certain objective(s) satisfying a number of conflicting constraints, as the selection of the optimal alternatives with a balancing act of a set of conflicting criteria, thus, falling in the domain of multi-criteria decision making (MCDM) problems. Traditionally, most of the MSS problems have been considered from technical and economical perspectives; but, increasing concerns by scientific societies and /or governments’ statutory regulations (e.g., Clean Air Act (1970), Resource Conservation and Recovery Act (1976) and Toxic Substance Control Act (1976)) on the environmental impacts of hazardous wastes produced by such systems during the operational or maintenance phases have forced organizations to devise and adopt more viable and sustainable solutions.