Job Scheduling Optimization in

Manufacturing System Using Genetic

Algorithm

Session 2020-2024

Group Members
Muneeb Ul Hassan 20-IE-02
Hasnat Ali Khalid 20-IE-20
Mustaneer Mehdi 20-IE-22
Jawad Ahmed Rana 20-IE-32
Raza Abbas Minhas 20-IE-46
Advisor
Engr. Muhammad Usman

Department of Industrial Engineering

University of Engineering and Technology
Taxila

1
 Abstract
This study focuses on optimizing job scheduling in manufacturing systems using a Genetic
Algorithm (GA). The objective is to minimize the makespan, which represents the total time
taken to complete all jobs, while considering constraints such as machine availability, job
prerequisites, and lot sizes. The GA iteratively evolves a population of candidate schedules to
find the most efficient arrangement of jobs on machines. The study involves decoding
individuals into schedules, evaluating the makespan as the objective function, and visualizing
the best solution using a Gantt chart. Through Python implementation, the study demonstrates
significant reductions in makespan, indicating improved efficiency in job scheduling for
manufacturing systems.

2
Contents
Chapter 1....................................................................................................................................5
Introduction................................................................................................................................5
1.1 Introduction.................................................................................................................5
1.2 Scheduling in Manufacturing:..........................................................................................6
Chapter 2....................................................................................................................................7
Case Study..................................................................................................................................7
2.1 Problem Statement...........................................................................................................7
2.2 Objectives.........................................................................................................................7
2.3 Significance of Study.......................................................................................................7
2.4 Data Collection and Preprocessing..................................................................................8
2.5 Methodological Approach................................................................................................8
2.6 Implementation Details....................................................................................................8
2.7 Evaluation Metrics...........................................................................................................8
2.8 Discussion of Results.......................................................................................................8
2.9 Limitations and Challenges..............................................................................................8
2.10 Conclusion......................................................................................................................8
Chapter 3....................................................................................................................................9
Methodology..............................................................................................................................9
3.1 Introduction......................................................................................................................9
3.2 Data Collection.................................................................................................................9
3.3 Input Code......................................................................................................................10
3.4 Approach of Genetic Algorithm (GA)...........................................................................11
3.5 Preprocessing.................................................................................................................12
3.6 Implementation Details..................................................................................................12
3.7 Evaluation of Scheduling Solutions...............................................................................12
3.8 Sensitivity Analysis and Parameter Tuning...................................................................13
Chapter 4..................................................................................................................................14
Results and Discussion.............................................................................................................14
4.1 Introduction....................................................................................................................14
4.2 Analysis of Optimization Process..................................................................................14
4.3 Evaluation of Scheduling Solutions...............................................................................14
4.4 Discussion......................................................................................................................15
4.5 Limitations and Future Directions.................................................................................15
4.6 Conclusion......................................................................................................................16

3
Chapter 5..................................................................................................................................17
Conclusions..............................................................................................................................17
5.1 Summary of Findings.....................................................................................................17
5.2 Contributions to Knowledge..........................................................................................17
5.3 Implications for Practice................................................................................................17
5.4 Future Research Directions............................................................................................17
5.5 Conclusion......................................................................................................................18

4
 Chapter 1
Introduction
1.1 Introduction
In modern manufacturing systems, efficient job scheduling plays a critical
role in maximizing productivity and minimizing costs. Job scheduling
involves assigning tasks to available resources, such as machines or
workers, while considering various constraints like processing times,
machine availability, and job dependencies. Traditional scheduling
methods often struggle to cope with the complexities of real-world
manufacturing environments, leading to suboptimal schedules and
inefficient resource utilization.
In contemporary manufacturing systems, the efficient allocation of
resources is crucial for meeting production targets, maintaining
competitiveness, and ensuring customer satisfaction. Job scheduling, a
fundamental aspect of production planning and control, involves assigning
tasks to available resources while adhering to various constraints and
objectives. The complexity of job scheduling arises from the dynamic
nature of manufacturing environments, which are characterized by:
Dynamic Workloads: Manufacturing systems often deal with fluctuating
demand, leading to varying workloads that necessitate flexible scheduling
strategies.
Resource Constraints: Limited availability of resources such as
machines, materials, and skilled labor poses significant challenges in
scheduling tasks to optimize utilization and minimize idle time.
Complex Dependencies: Certain tasks have dependencies on preceding
operations, equipment setup times, or specific sequences, requiring
careful coordination to avoid disruptions and delays.
Multi-objective Optimization: Job scheduling typically involves multiple
conflicting objectives, such as minimizing makespan, reducing setup
times, and balancing workload distribution, necessitating sophisticated
optimization techniques.
Traditional scheduling approaches, including rule-based heuristics and
mathematical programming models, often struggle to cope with the
complexity and uncertainty inherent in real-world manufacturing
environments. As a result, suboptimal schedules may lead to
inefficiencies, production delays, and increased costs.
To address these challenges, researchers and practitioners have turned to
advanced optimization techniques, such as Genetic Algorithms (GAs),
inspired by principles of natural selection and evolution. GAs offer a robust
and versatile approach to solving complex optimization problems by
5
mimicking the process of natural evolution, wherein candidate solutions
(individuals) undergo selection, crossover, mutation, and reproduction to
generate improved offspring over successive generations.
In the context of job scheduling optimization, GAs offer several
advantages:
Exploration of Solution Space: GAs explore a wide range of potential
solutions, enabling the discovery of diverse and innovative schedules that
may not be apparent through traditional methods.
Adaptability to Constraints: GAs can accommodate various constraints
and objectives, making them suitable for addressing the multifaceted
nature of job scheduling problems in manufacturing systems.
Parallelization and Scalability: GAs can be parallelized and scaled to
handle large-scale scheduling problems efficiently, leveraging advances in
computing technology.
By leveraging GAs and other evolutionary optimization techniques,
researchers aim to develop robust and adaptive scheduling algorithms
capable of generating high-quality schedules that optimize resource
utilization, minimize production lead times, and enhance overall
operational efficiency in manufacturing systems.

1.2 Scheduling in Manufacturing:

Scheduling in manufacturing refers to the process of determining the
sequence and timing of production activities to efficiently utilize available
resources (such as machines, materials, and labor) while satisfying
various constraints and objectives. In manufacturing environments,
scheduling decisions directly impact production efficiency, lead times,
resource utilization, and ultimately, the overall performance of the
manufacturing system.
The key objectives of scheduling in manufacturing include:
Minimizing Makespan: The total time required to complete all
production tasks, also known as makespan, is a critical performance
metric. Minimizing makespan helps reduce production lead times and
ensures timely delivery of products to customers.
Optimizing Resource Utilization: Efficient scheduling aims to maximize
the utilization of available resources, such as machines and labor, while
minimizing idle time and minimizing production bottlenecks.
Balancing Workload: Distributing workload evenly across machines and
workers helps prevent overloading of resources and ensures smooth
production flow.
Meeting Due Dates: Scheduling should prioritize tasks to meet
customer due dates and minimize order lateness or earliness.

6
Minimizing Setup and Changeover Times: Efficient scheduling
involves reducing setup and changeover times between different
production runs, enabling quick transitions and flexible production.
Given the complexity of manufacturing environments and the multitude of
factors to consider, scheduling in manufacturing often poses a challenging
optimization problem. Traditional scheduling methods, such as rule-based
heuristics and mathematical programming models, may struggle to
effectively address the dynamic and uncertain nature of manufacturing
operations.

Chapter 2
Case Study
In this chapter, we present a case study focusing on the application of Genetic Algorithm
(GA) for job scheduling optimization in a manufacturing system. The case study provides a
practical illustration of how GAs can be employed to address complex scheduling challenges
and optimize production processes.

2.1 Problem Statement

The complexity of job scheduling in manufacturing systems arises due to several factors,
including:
Machine Availability: Limited availability of machines can lead to bottlenecks and idle
time, affecting overall production efficiency.
Job Prerequisites: Certain jobs have dependencies, requiring specific tasks to be completed
before they can start, and adding complexity to the scheduling process.
Lot Sizes: Varying lot sizes further complicate scheduling, as different production runs may
require adjustments in scheduling priorities.
Addressing these challenges is crucial for improving overall operational efficiency and
reducing production costs in manufacturing systems.

2.2 Objectives
The primary objective of this study is to optimize job scheduling in manufacturing systems
using a Genetic Algorithm (GA). Specifically, the study aims to:
 Minimize the makespan, representing the total time taken to complete all jobs, by
efficiently allocating tasks to available resources.
 Consider constraints such as machine availability, job prerequisites, and lot sizes to
generate feasible and optimized schedules.
 Utilize the GA to iteratively evolve a population of candidate schedules, aiming to
find the most efficient arrangement of jobs on machines.

2.3 Significance of Study

7
Efficient job scheduling can lead to several benefits for manufacturing systems, including:
Improved Production Efficiency: Optimized schedules ensure smoother workflow,
minimizing idle time and maximizing resource utilization.
Reduced Costs: By minimizing makespan and optimizing resource allocation, manufacturing
systems can reduce production costs and improve profitability.
Enhanced Competitiveness: Efficient scheduling allows manufacturers to meet customer
demands more effectively, enhancing their competitiveness in the market.

2.4 Data Collection and Preprocessing

The case study involves collecting relevant data related to job
characteristics, machine specifications, lot sizes, processing times, and
other parameters essential for scheduling optimization. Data
preprocessing techniques may be employed to clean, organize, and
transform the raw data into a suitable format for input into the scheduling
algorithm. Additionally, historical production data and performance
metrics may be analyzed to gain insights into system dynamics and
inform scheduling decisions.

2.5 Methodological Approach

The methodological approach adopted in the case study revolves around
the implementation of a Genetic Algorithm for job scheduling
optimization. The GA operates iteratively, evolving a population of
candidate schedules through selection, crossover, mutation, and
reproduction. The fitness of each schedule is evaluated based on
predefined objectives, such as minimizing makespan and satisfying
constraints such as machine availability and job dependencies.

2.6 Implementation Details

The implementation of the GA involves encoding scheduling solutions into
a genetic representation, defining appropriate genetic operators
(selection, crossover, mutation), and specifying parameters such as
population size, crossover probability, and mutation rate. The GA
iteratively generates and evaluates schedules, guiding the search towards
optimal or near-optimal solutions.

2.7 Evaluation Metrics

To assess the performance of the scheduling algorithm, various evaluation
metrics may be employed, including makespan, resource utilization,
production lead time, and schedule stability. These metrics provide
insights into the effectiveness of the scheduling approach and help
identify areas for improvement.

8
2.8 Discussion of Results
The results of the case study are presented and discussed, highlighting
the effectiveness of the GA in generating optimized schedules and
improving production efficiency. The discussion may include comparisons
with baseline scheduling methods, analysis of solution quality, sensitivity
to parameter settings, and implications for real-world manufacturing
operations.

2.9 Limitations and Challenges

Despite its effectiveness, the scheduling approach may have limitations
and face challenges in certain scenarios. Factors such as computational
complexity, scalability, and robustness to uncertainties may impact the
applicability and performance of the GA-based scheduling algorithm.
Understanding these limitations is essential for interpreting results and
guiding future research directions.

2.10 Conclusion
The chapter concludes with a summary of key findings from the case
study, emphasizing the potential of GAs for job scheduling optimization in
manufacturing systems. The insights gained from the case study
contribute to the body of knowledge in scheduling theory and practices.

Chapter 3
Methodology
3.1 Introduction
This chapter outlines the methodology employed in the case study to optimize job scheduling
in a manufacturing system using a Genetic Algorithm (GA). The methodology encompasses
various stages, including data collection, preprocessing, implementation of the GA, and
evaluation of scheduling solutions.

3.2 Data Collection

Data collection involves gathering relevant information about the manufacturing system,
including job characteristics, machine specifications, lot sizes, processing times, and any
other parameters necessary for scheduling optimization. The input data for the GA is obtained
from sources such as production records, equipment specifications, and operational
databases. In the present study, the input data is derived from the "Dataset.xlsx" file, which
contains information about jobs, lots, operations, machines, and lot sizes.

9
In above data, there is lot size of 18 observations with first six have lot size of 1, next six
have lot size of 2 and remaining have lot size of 3. And in the next column there are operation
number and then in the next column there are number of machines assigned to each lot size
an and in 4th column there is an process time of every lot.
And the three lots have there size as lot size 1, 2 and 3 next in the separate columns.

10
3.3 Input Code

11
3.4 Approach of Genetic Algorithm (GA)
The Genetic Algorithm (GA) serves as the core optimization technique for job scheduling in
the manufacturing system. The GA operates iteratively, simulating the process of natural
selection to evolve a population of candidate schedules towards optimal or near-optimal
solutions. Key components of the GA include:
Representation: Scheduling solutions are encoded as individuals in a population, with each
individual representing a potential schedule of jobs on machines.
Fitness Evaluation: The fitness of each schedule is determined based on predefined
objectives, such as minimizing makespan, while considering constraints such as machine
availability and job dependencies.
Genetic Operators: Selection, crossover, mutation, and reproduction are employed to
manipulate the population and generate new schedules. These operators enable exploration of
the solution space and facilitate convergence towards better solutions.

3.5 Preprocessing
Preprocessing of input data involves cleaning, organizing, and transforming raw data into a
suitable format for input into the GA. In the context of job scheduling, preprocessing tasks
may include removing duplicates, handling missing values, sorting jobs based on priorities,
and adjusting parameters such as lot sizes and processing times. Additionally, preprocessing
techniques may be employed to encode constraints and objectives into the genetic
representation of scheduling solutions.

3.6 Implementation Details

The implementation of the GA entails defining the genetic representation, specifying genetic
operators, setting algorithm parameters, and integrating the scheduling logic into the GA
framework. The GA iteratively generates and evaluates schedules, guiding the search towards
optimal or near-optimal solutions. Implementation details include:

12
Genetic Representation: Encoding of scheduling solutions as lists of job indices,
representing the sequence of jobs to be performed on machines.
Genetic Operators: Selection, crossover, and mutation operators tailored to the scheduling
problem, facilitating the exploration of solution space and generation of diverse schedules.
Algorithm Parameters: Specification of parameters such as population size, crossover
probability, mutation rate, and termination criteria to control the behavior and convergence of
the GA.
The implementation details are realized through the Python code provided, which utilizes the
DEAP (Distributed Evolutionary Algorithms in Python) framework for implementing the GA
and solving the job scheduling optimization problem.

3.7 Evaluation of Scheduling Solutions

Following the implementation of the Genetic Algorithm (GA) for job scheduling
optimization, the next step involves evaluating the quality and effectiveness of the generated
scheduling solutions. Evaluation metrics are employed to assess various aspects of the
schedules, including performance, efficiency, and adherence to constraints. Common
evaluation metrics used in scheduling optimization include:
Makespan: The total time taken to complete all jobs in the schedule. Minimizing makespan
is a primary objective in job scheduling optimization, as it reflects the overall efficiency of
the production process.
Resource Utilization: The extent to which manufacturing resources, such as machines and
operators, are utilized in the schedule. Efficient resource utilization ensures optimal use of
available capacity and minimizes idle time.
Production Lead Time: The time elapsed from the initiation to the completion of a job.
Minimizing production lead time is crucial for meeting customer deadlines and enhancing
responsiveness in the production process.
Schedule Stability: The consistency and reliability of the schedule over time. A stable
schedule ensures smooth and predictable operations, reducing the likelihood of disruptions
and delays.
Evaluation of scheduling solutions may involve comparing the performance of the GA-
generated schedules with baseline schedules or existing scheduling methods. Statistical
analysis, visualization techniques, and sensitivity analysis may be employed to gain insights
into the strengths and weaknesses of the scheduling approach and identify opportunities for
improvement.

3.8 Sensitivity Analysis and Parameter Tuning

To ensure robustness and effectiveness, sensitivity analysis and parameter tuning may be
performed to investigate the impact of different algorithm parameters on the performance of
the GA. Sensitivity analysis involves systematically varying one or more parameters while
keeping others constant and observing the resulting changes in solution quality. This helps
identify optimal parameter configurations that lead to improved scheduling outcomes.

13
Parameter tuning involves adjusting algorithm parameters, such as population size, crossover
probability, and mutation rate, to optimize the performance of the GA. Techniques such as
grid search, random search, or heuristic approaches may be employed to systematically
explore the parameter space and identify configurations that yield better results.

Chapter 4
Results and Discussion
4.1 Introduction
This chapter presents the results obtained from the application of the Genetic Algorithm (GA)
to the problem of job scheduling in manufacturing. The discussion delves into the
performance of the GA-based optimization approach and analyzes the generated scheduling
solutions in terms of their efficiency, effectiveness, and adherence to predefined objectives.

4.2 Analysis of Optimization Process

The optimization process involved running the GA for a specified number of generations,
with each generation producing a new population of candidate scheduling solutions. The
table provided showcases the evolution of the optimization process, indicating the number of
generations, the average makespan of each generation, and other relevant statistics such as
the minimum and maximum makespan observed.

14
4.3 Evaluation of Scheduling Solutions
The best individual solution generated by the GA is presented, along with the corresponding
scheduling details, including the assignment of lots to operations and machines, as well as the
start and finish times for each operation. The makespan of the best solution, which represents
the total time taken to complete all jobs in the schedule, is highlighted.
Additionally, a graphical representation of the optimization process is provided, depicting the
trend of the minimum and average makespan values across generations. This graphical
analysis offers insights into the convergence behavior of the GA and illustrates how the
makespan improves over successive generations.

15
4.4 Discussion
The results are discussed in the context of the objectives outlined in the study. The achieved
makespan value is compared against baseline makespan values or benchmarks, providing a
basis for assessing the effectiveness of the GA-based optimization approach.
Key factors influencing the performance of the GA, such as parameter settings, population
size, and convergence criteria, are analyzed. Insights are drawn regarding the impact of these
factors on solution quality and computational efficiency.
Furthermore, the scheduling solutions are evaluated with respect to their practical
applicability and feasibility in real-world manufacturing environments. Considerations such
as resource constraints, production priorities, and scheduling flexibility are addressed,
highlighting the relevance of the GA-generated schedules for optimizing manufacturing
operations.

4.5 Limitations and Future Directions

The chapter concludes with a discussion of limitations encountered during the study and
potential avenues for future research. Limitations may include constraints inherent to the GA
approach, such as scalability issues or sensitivity to parameter settings, as well as challenges
specific to the manufacturing context, such as dynamic job arrivals or uncertain processing
times.
Future research directions may involve exploring alternative optimization techniques,
integrating additional constraints or objectives into the scheduling model, or extending the
study to different manufacturing scenarios or industries. Suggestions for further improving
the efficiency and effectiveness of the scheduling optimization process are provided, paving
the way for continued advancements in manufacturing scheduling practices.

4.6 Conclusion
In summary, the results and discussion chapter provides a comprehensive analysis of the GA-
based optimization approach for job scheduling in manufacturing. By presenting detailed
results, discussing key findings, and identifying avenues for future research, the chapter
contributes to a deeper understanding of scheduling optimization methods and their
applications in enhancing manufacturing operations.

16
 Chapter 5
Conclusions
5.1 Summary of Findings
In this study, we investigated the application of Genetic Algorithm (GA) for optimizing job
scheduling in manufacturing operations. The research aimed to minimize the makespan,
which represents the total time required to complete all jobs, thereby enhancing the efficiency
and productivity of manufacturing processes. Through the implementation and evaluation of
the GA-based optimization approach, several key findings emerged:
The Genetic Algorithm demonstrated effectiveness in generating scheduling solutions that
significantly reduce the makespan compared to conventional methods.
The optimization process exhibited convergence behavior, with the makespan steadily
improving across successive generations of the GA.
The scheduling solutions produced by the GA were evaluated for their practical applicability
and feasibility in real-world manufacturing environments, demonstrating their potential to
enhance production efficiency and meet scheduling objectives.

5.2 Contributions to Knowledge

This study contributes to the body of knowledge on scheduling optimization in manufacturing
by:

17
Demonstrating the effectiveness of Genetic Algorithm as a viable approach for addressing
complex scheduling problems.
Providing insights into the performance and behavior of the GA-based optimization process,
including its convergence characteristics and solution quality.
Highlighting the practical implications of optimized scheduling solutions for improving
manufacturing operations, such as minimizing production lead times and maximizing
resource utilization.

5.3 Implications for Practice

The findings of this study have several implications for practitioners in the manufacturing
industry:
Adoption of Genetic Algorithm-based scheduling optimization techniques can lead to
significant improvements in production efficiency and resource utilization.
Real-time scheduling systems leveraging GA algorithms can facilitate dynamic adaptation to
changing production environments and demand fluctuations.
Integration of scheduling optimization tools into existing manufacturing software platforms
can enhance decision-making capabilities and support agile production planning.

5.4 Future Research Directions

While this study provides valuable insights into the application of Genetic Algorithm for job
scheduling optimization in manufacturing, several avenues for future research exist.
Exploration of hybrid optimization approaches combining Genetic Algorithm with other met
heuristic or machine learning techniques to further improve solution quality and convergence
speed.
Investigation of multi-objective optimization frameworks that consider additional
performance criteria such as energy consumption, machine utilization, and job tardiness.
Evaluation of the scalability and robustness of GA-based scheduling optimization methods
for large-scale manufacturing systems and complex production environments.
Integration of real-time data analytics and predictive modeling techniques to enhance the
adaptability and responsiveness of scheduling systems to dynamic production conditions.

5.5 Conclusion
In conclusion, this study underscores the potential of Genetic Algorithm as a powerful
optimization tool for addressing job scheduling challenges in manufacturing. By leveraging
the strengths of evolutionary computation, GA-based approaches offer practical solutions for
improving production efficiency, reducing lead times, and enhancing overall operational
performance. Moving forward, continued research and development efforts in this domain
hold promise for advancing scheduling optimization practices and driving innovation in
manufacturing processes.

18
19