Welcome to the Production and Operations Analytics (POA) Research Laboratory at the University of Windsor's Department of Mechanical, Automotive & Materials Engineering (MAME).
Research Program:
Our research program focuses on advancing Production and Operations Management (POM) by developing intelligent planning functions and logical enablers for Smart Manufacturing (SM) and Industry 4.0, applying both predictive and prescriptive analytics. Recognizing the critical need to balance technological advancement with ecological and economic viability, our work firmly embeds sustainability, flexibility, and predictability into the core of operations modeling. By integrating Artificial Intelligence (AI), Machine Learning (ML), and advanced operations research optimization, we solve complex planning problems across manufacturing floors, global supply chains, power grids, and healthcare facilities.
We primarily focus on developing the logical enablers for Industry 4.0. A primary thrust of our lab is operationalizing highly flexible paradigms, specifically Hybrid Manufacturing (HM), and ensuring manufacturing systems can autonomously adapt to disruptions.
Computer Aided Process Planning (CAPP): We employ different Generative-AI architectures to process CAD data structures, undergoing both Features Recognition and macro-CAPP. For synthesis of corpus, we optimize the sequencing and grouping of manufacturing features for HM by hybridizing optimization models (like Mixed Integer Programming) with ML techniques: Logical Analysis of Data.
At the micro level, we utilize ML, Deep Learning, and optimization to predict and minimize the environmental footprint of machining parameters, actively targeting reductions in energy consumption and greenhouse gas emissions.
Adaptive Lot-Sizing and Closed-Loop Scheduling: To manage the stochastic demand inherent in mass customization, we develop multi-period, multi-objective Stochastic Programming and Robust Optimization (RO) models for HM. Furthermore, we design closed-loop integrated CAPP and scheduling systems that dynamically reconfigure resource assignments and sequences based on real-time data fed from shop-floor Digital Twins.
Dynamic Layouts and Resilient Supply Chains: To support true factory transformability, we address the dynamic Facility Layout Problem using Dynamic Programming and RL, allowing manufacturing fractals to optimize their self-reconfiguration strategies. Beyond the factory wall, we tackle the Facility Location and Allocation Problem by leveraging ML to quantify political risks in emerging economies, integrating these predictions into Integer Linear Programming models to optimize global supply chains.
Extensions: Power Grids & Healthcare
The robust POM mathematical models and AI hybrid algorithms developed in our lab are highly adaptable and are being extended to solve critical logistical challenges in other vital sectors.
Power Grid Planning for Electric Vehicles (EVs): We model the siting and sizing of Distributed Generation (DG) and EV charging infrastructure to accommodate the acute loads introduced by EV adoption.
To mitigate the intermittency of renewable DG and regulate grid voltage, we formulate Vehicle Routing Problems (VRP) and scheduling models to actively deploy fleets of Mobile Energy Storage Systems (MESSs) and mobile STATCOMs precisely where the grid needs them.
Healthcare Operations Management: We are developing a comprehensive Decision Support System to optimize operating room planning across strategic, tactical, and operational horizons. This includes utilizing RO and strategy pattern hyper-heuristics to manage multi-mode appointment scheduling under the persistent uncertainties of patient arrivals and no-shows.