Optimization of production line balancing and line performance under variable demand using hybrid metaheuristics in a discrete-event simulation environment

Project Abstract

This study presents a novel framework for optimizing production line balancing and overall line performance under variable demand by integrating hybrid metaheuristics within a discrete-event simulation (DES) environment. The core objective is to minimize makespan, balance workload, reduce work-in-process, and improve key performance indicators such as throughput, cycle time, and capital utilization, while maintaining robustness to demand fluctuations and stochastic processing times. The methodology combines the global search capabilities of metaheuristics (genetic algorithms, simulated annealing, and particle swarm optimization) with problem-specific heuristics and local search strategies to generate high-quality line configurations, task assignments, and sequencing policies. The DES component models the production line as a sequence of workstations with finite buffers, setup times, machine reliability, and variable processing times drawn from empirical distributions, enabling realistic evaluation of line performance under different demand scenarios. A modular optimization framework is proposed to simultaneously address line balancing, bottleneck mitigation, and line configuration adaptation in response to demand variability. The framework defines a multi-objective optimization problem that incorporates operational constraints such as precedence relations, workstation capabilities, changeover requirements, and maintenance windows. To handle the stochastic nature of manufacturing, scenario-based and time-varying demand inputs are embedded within the DES runs, and performance is assessed using statistical measures including mean makespan, standard deviation, cycle time distribution, throughput rate, and service level. The hybrid algorithm orchestrates exploration and exploitation across two layers (i) a global search layer that redistributes tasks among stations and sequences operations to reduce idle times and balance workload, and (ii) a local refinement layer that fine-tunes task assignments and routing within nearby neighborhoods to improve stability and predictability of the line under demand shocks. The research contributes methodological advances by (a) integrating DES with hybrid metaheuristics to capture dynamic and stochastic effects in a single evaluative loop, (b) proposing a robust objective function that blends economic and operational performance, and (c) presenting a scalable optimization scheme suitable for large-scale lines with hundreds of tasks and multiple product mixes. Data from a mid- to large-scale manufacturing setting, including task times, failure rates, setup durations, and demand profiles, are used to calibrate and validate the model. Comparative analyses against conventional line balancing heuristics and pure metaheuristic approaches demonstrate superior performance in terms of reduced average cycle time, lower variance in workstation utilization, and improved service levels without compromising throughput. Sensitivity analyses examine the impact of demand volatility, buffer capacities, and maintenance schedules on optimization outcomes, guiding practitioners in choosing appropriate risk-aware settings. The study also develops practical implementation guidelines, including step-by-step integration strategies for industrial environments, computational considerations, and visualization tools to monitor real-time line performance and adaptive scheduling in response to demand fluctuations. Overall, the research advances the state-of-the-art in production system optimization by delivering an effective, adaptable, and scalable solution for balancing efficiency and resilience in the face of variable demand.

Project Overview

What This Project Is About

---

This project looks at how to organize a production line so work moves smoothly without delays, even when demand changes. It combines ideas from two approaches—hybrid metaheuristics (smart, flexible problem-solving methods) and discrete-event simulation (a way to model how a factory operates over time). The goal is to find a balanced sequence of tasks and a way to schedule work that keeps production efficient and responsive to fluctuations in orders.

---

---

The Problem It Addresses

---

Many factories struggle with bottlenecks, uneven workloads, and idle times when demand varies. Traditional planning methods may not adapt quickly, causing longer lead times and higher costs. This project tackles how to automatically re-balance lines and improve overall performance under changing demand, with potential benefits for productivity and customer satisfaction.

---

---

Objectives of the Project

---

1. Understand the fundamentals of line balancing and throughput in a factory.
2. Explore how hybrid metaheuristics can help find good production schedules.
3. Develop a simple discrete-event simulation model of a production line.
4. Integrate the two approaches to test different line configurations under variable demand.
5. Evaluate performance using clear metrics such as cycle time and utilization.

---

---

What You Will Do Step by Step

---

1. Review basic concepts of line balancing and production simulation.
2. Build a small, representative factory model with stations and tasks.
3. Implement a hybrid search method to optimize task assignments and timings.
4. Run simulations under different demand scenarios and collect results.
5. Compare optimized plans to a baseline to measure improvements.
6. Analyze sensitivity to parameters and provide practical recommendations.

---

---

Expected Outcome

Expected outcomes include a validated approach that improves line balance and reduces idle time under variable demand, with clear guidance for practitioners on implementing such methods in real plants.