# Research Activities

## SUMMARY OF RESEARCH ACTIVITIES AT UNIVERSITY OF GUELPH

### Integrated Production and Manpower Scheduling

The research problem identified in a Local Industry involves the scheduling of very large number of orders in its Toronto manufacturing facility. The system consists of a large number of stations that process many batches of products. Some stations have identical parallel machines that share limited number of tools. There are also limited number of cross trained operators. In some processing stations, an operator can be assigned to tend more than one machine at a time. The products need assembly operations where the components and subassemblies including the final assemblies need to pass through sequences of processing stations. Some processing and assembly operations (semi-finished products) are also required to be sent to an outside manufacturer where the products are returned after a specified lead time for further in-house processing. The jobs have to be scheduled to meet due dates and at the same time the load across the parallel machines and among the operators need to be balanced. The work-in-process inventory has to be maintained at the lowest possible. This scheduling problem is different from standard problems that are well documented in literature. To solve this new scheduling problem, we applied the event-scheduling/time-advance algorithm from the theory of discrete event systems simulation.

### Plastic Mold Station scheduling in Automotive Parts Paint Facility

This research is in collaboration with a local industry and funded by NSERC-Engage at a level of 21,000 CAD. Preliminary studies were conducted during Fall 13 semester to define the scope of the problem and roadmap to its solution procedure. The problem identified in the collaborating industry involves optimizing the placement of different vehicle exterior components with varying styles and color in a complex painting and assembly facility. The goal is to identify the most efficient schedule to meet customer demand and flow part styles to the appropriate warehouse spaces. The constraints of the process are as follows: (1) The schedule must meet customer demand, (2) Parts are placed on carriers which have a fixed capacity - a given part must be put on its unique carrier, (3) Certain carriers may not be placed side by side due to physical constraints of the paint line, (4) Swapping carriers out from one round to the next is possible but is highly resource expensive and so be minimized, (5) Parts may be shuttled via an overhead carrier to the a sequence center if room allows, (6) The sequence center is a predefined space by a paint line scheduler, (7) Alternatively, parts may be loaded into racks and moved from the paint line to the warehouse by a tow motor driver, (9) Mixing racks by style or color is prohibited therefore, it is desirable to fill a rack completely, (10) Tow motor drivers can carry four racks at one time to the warehouse and return with four empty racks to feed the paint line with a total round trip of ten minutes, (11) Four empty racks can be stored line side at any time, (12) Total time to rack parts is seven times greater than to feed the sequence center, (13) Sequence center space is a fraction the size of the warehouse, and finally (14) Both forms of unloading the paint line must keep pace with the speed of the paint line. As a preliminary study, a scheduling algorithm was developed for planning the operations of many molding stations that produce more than 28,000 plastic parts a day. This parts are bumpers and other exterior automotive components that will be painted on three automated painting lines.

### Scheduling in a local food Processing Industry

The vast majority of algorithms appeared in literature are of limited use to companies with specific problems. Real-world scheduling problems are extremely different from research problems and quite often they are very complicated. A research project I encountered in a local industry is a typical example of these types of scheduling problem. A highly paced upstream continues process of a constant speed is feeding hundreds of intermittent downstream operations interconnected by network of conveyors. The scheduling of the downstream operations calls for a balance of material flow at all processing stations and a timely demand satisfaction of products.

### An integrated approach in the design and operation of non-conventional manufacturing systems

There is an emerging consensus that traditional layout configurations do not meet the needs of multi-product enterprises working in volatile environment. For these types of enterprises, there is a need for new generation of factory layouts that are more flexible, modular, and easy to reconfigure. I am conducting research in this area since May 2011 to gain in-depth undersigning to rapid modeling and simulation of non-conventional manufacturing systems to answer questions related to: changes in product mix and volume; relocation, addition and deletion of machine tools; reconfiguration of material handling system; operator assignment, shift-scheduling and capacity adjustment; and product lot sizing and scheduling. Main contribution of this research will be in developing: (1) Models, algorithms and computerized tools to design and optimize non-conventional manufacturing systems such as cellular, fractal and reconfigurable manufacturing systems; (2) Manufacturing systems performance metrics tailored to the design criteria of these non-conventional manufacturing systems; and (3) Simulation model to evaluate several reconfiguration decisions in reconfigurable manufacturing systems. This research is funded by NSERC at a level of 100,000 CND for a period of from 2011 to 2016. Additional funds worth of 40,000 through university start-up fund and 40,200 CAD through NSERC-RTI have been secured to support this research.

### Manufacturing systems design based on distributed layout

This is a research effort in advancing theory in manufacturing systems design and operations. The design and operation of production systems in the current era of global competition is becoming a difficult task and very complex. Modeling and optimization of such complex systems is of paramount importance in achieving competitive advantages. In this research, I developed a new mathematical model that integrates layout configuration and production planning in the design of dynamic distributed layouts. These types of layout are emerging as a remedy to the challenges faced by manufacturing systems producing multiple components and working in today's highly volatile environments. The developed model incorporates a number of important manufacturing attributes such as demand fluctuation, system reconfiguration, lot splitting, work load balancing, alternative routings, machine capability and tooling requirements. In addition, the model allows the optimization of several cost elements in an integrated manner. These include material handling, machine relocation, setup, inventory carrying, in-house production and subcontracting costs. The developed model is difficult to solve using off-the-shelf optimization packages. To this end, I developed multiple search path simulated annealing algorithms. Computational performance of the proposed heuristic search method is very encouraging based the results of the testing problems. I am currently working to further enhancing the developed model to account for uncertainties in product demand and mix. Moreover, I am also working in developing a scheduling algorithm for distributed layout based manufacturing systems where existing research is very limited. One PhD student is involved in this research and an article has been submitted to the International Journal of Production Economics.

### Linear programming assisted genetic algorithm for job shop scheduling with lot streaming

I began this research in summer 2012. The hybridization of metaheuristics with other techniques for optimization has been one of the most interesting trends in recent years. The focus of research on metaheuristics has also notably shifted from an algorithm-oriented point of view to a problem-oriented point of view. Many researchers focus on solving a problem at hand as best as possible rather than promoting a certain metaheuristic. This has led researchers to attempt combining different algorithmic components in order to design algorithms that are more powerful than those resulting from the implementations of a pure metaheuristic. In this research, I developed a linear programming assisted genetic algorithm for solving a flexible job-shop scheduling problem with lot streaming. The genetic algorithm searches over both discrete and continues variables in the problem solution space. A linear programming is used to further refine promising solutions in the initial population and during the genetic search process by determining the optimal values of continuous variables corresponding to the values of the integer variables of these promising solutions. Detail numerical studies were conducted and the results showed that the proposed algorithm outperforms previously published results. One master student was involved in this research and the student defended his thesis on Feb 25, 2014. Currently, an article is under predation and will be submitted to a journal within summer 2014.

### Product cost estimation

This is a collaborative research between me, Dr. Nadia Bhuiyan (Concordia University) and Dr. Addis Salam (a senior cost engineer at Bombardier Aerospace). The following interrelated researches were conducted between Winter 2010 to Summer 2012.

New Approach for Preliminary Product Costing using Data Envelopment Analysis: In this research, we developed a new cost estimation technique based on Data Envelopment Analysis (DEA). In operations research and economics literature, DEA has been used as tools for the estimation of relative efficiencies and ranking of decision making units (organizational bodies). In this research, we extended the use of DEA beyond this traditional application to a new cost estimation tool. This work is published recently in the International Journal of Industrial Engineering computations.

A Case Study on Target Cost Estimation using Back-Propagation and Genetic Algorithm Trained Neural Networks: Cost estimation of new products has always been difficult as only few attributes will be known. In these situations, parametric methods are commonly used using a priori determined cost function where parameters are evaluated from historical data. Neural networks, in contrast, are non-parametric, i.e. they attempt to fit curves without being provided a predetermined function. In this paper, this property of neural networks is used to investigate their applicability for cost estimation of certain major aircraft subassemblies. The study is conducted in collaboration with an aerospace company located in Montreal, Canada. Two neural network models, one trained by the gradient descent algorithm and the other by genetic algorithm, are considered and compared with one another. The study, using historical data, shows an example for which the neural network model trained by genetic algorithm is robust and fits well both the training and validation data sets. The result of this research has been recently published in the Journal of Cost Analysis and Parametric.

A case study to estimate costs using neural networks and regression based models: Bombardier Aerospace’s high performance aircrafts and services set the utmost standard for the Aerospace industry. A case study in collaboration with Bombardier Aerospace (Dr. Nadia Bhuiyan of Concordia as a principal investigator) is conducted in order to estimate the target cost of a landing gear. More precisely, the study uses both parametric model and neural network models to estimate the cost of main landing gears, a major aircraft commodity. A comparative analysis between the parametric based model and those upon neural networks model was considered in order to determine the most accurate method to predict the cost of a main landing gear. The result of this research appeared recently in the Decision Science Letters.

### Transportation of Multiple Petroleum Products through a Pipeline System

This research was a collaborative research between me, Dr. Mingyuan Chen (Concordia University) and two professor at Complutense University (SPAIN). My involvement in this research started while I was at Concordia and was continued during Winter and Summer 2010 (after I joined University of Guelph). The research is about operation planning of multiproduct pipeline system. A multiproduct pipeline provides an economic way to transport large volumes of refined petroleum products over long distances. In such a pipeline, different grades of petroleum products are pumped back−to−back without any separation device between them. The sequence and lengths of such pumping runs must be carefully selected in order to satisfy several pipeline system constraints and minimizing operational costs. The production planning and scheduling of the products at the refinery must also be synchronized with the transportation. In this work, we developed a multi−period mixed integer nonlinear programming (MINLP) model for an optimal planning and scheduling of the production and transportation of multiple petroleum products from a refinery plant connected to several depots through a single pipeline system. A paper published on 2011 reporting this research.

## SUMMARY OF RESEARCH ACTIVITIES AT CONCORDIA UNIVERSITY WHEN I WAS A POSTDOC/RESEARCH ASSOCIATE

### Development of mathematical models and solution procedure for manufacturing systems scheduling

This research encompasses the development of very comprehensive mathematical models and solution procedures for scheduling job-shop and flow-shop based manufacturing systems. The developed mathematical scheduling models incorporate several pragmatic issues such as (1) the existence of alternative routing in processing parts, (2) lot streaming techniques to reduce production makespan, (3) sequence dependence of processing setup, (4) attached or detached natures of processing setups, (5) release dates of machine tools from previous schedules, among others. The developed solution procedures are based on meta-heuristic algorithms known as genetic algorithm and simulated annealing. The application of parallel computing has been also utilized in solving the proposed models. The results of this research are been published in very recognized international journals called (1) international journal of production research, and (2) international journal of advanced manufacturing technology.

### Decision support systems

During summer 2008, I worked on industrial project from Pratt & Whitney Canada (an aerospace company) in collaboration with Concordat University. I worked in developing a decision support system that takes bottleneck data of producer/supplier lines versus engine families and highlights the opportunity of trade-offs among engine families with the objective to minimize the number of engines that would be overdue by the end of the planning period under consideration. This research activity was supervised by Dr. Nadia Bhuiyan, P.ENG and Associate Professor at Concordia University.

## SUMMARY OF RESEARCH ACTIVITIES AT CONCORDIA UNIVERSITY WHEN I WAS A PHD STUDENT

### An integrated approach to the design of cellular manufacturing systems

As a PhD student I was conducting research in developing a comprehensive model and solution produced for a cellular (group technology) manufacturing system. The design of cellular manufacturing systems (CMS) involves many structural and operational issues. One of the important design steps is the formation of part families and machine cells. In this thesis, a comprehensive mathematical model for the design of CMS based on tooling requirements of the parts and tooling available on the machines was proposed. The model incorporates dynamic cell configuration, alternative routings, lot splitting, sequence of operations, multiple units of identical machines, machine capacity, workload balancing among cells, operation cost, cost of subcontracting part processing, tool consumption cost, setup cost, cell size limits, and machine adjacency constraints. To solve this model efficiently, a two-phase genetic-algorithm-based heuristic was developed. In the first phase, independent cells are formed which are relatively simple to generate. In the second phase, the solution found during the first phase is gradually improved to generate cells optimizing inter-cell movement and other cost terms of the model.

This research was further expanded to incorporate production planning by considering the impact of lot sizes on product quality. Production lot sizing models are often used to decide the best lot size to minimize operation cost, inventory cost, and setup cost. Cellular manufacturing analyses, on the other hand, mainly address how machines should be grouped and parts be produced. In this research endeavor, I developed a mathematical programming model following an integrated approach for cell configuration and lot sizing in a dynamic manufacturing environment. The model development also considers the impact of lot sizes on product quality. Solution of the mathematical model is to minimize both production and quality related costs. The proposed model, with nonlinear terms and integer variables, was difficult to be solved for real size problems efficiently due to its NP-complexity. To solve the model for practical purposes, a linear programming embedded genetic algorithm was developed. The algorithm searches over the integer variables and for each integer solution visited the corresponding values of the continuous variables are determined by solving a linear programming sub-problem using the simplex algorithm.