Abstract—In today’s world, improving efficiency and reducing costs in logistics services are critical challenges for organizations, as bottlenecks and process non-conformities can significantly disrupt performance and competitiveness. This research, using a process mining approach, aims to identify and analyze these issues in detail. Data related to logistics processes were collected from organizational management systems, and after cleaning, were analyzed using K-Prototype models for clustering, Random Forest Classifier for prediction, and Apriori for discovering association rules. The results indicate that process mining can effectively identify bottlenecks and process inconsistencies, providing innovative solutions to improve operational times, reduce costs, and enhance process alignment with designed models. This study highlights the importance and effectiveness of process mining in improving performance and increasing the competitiveness of logistics services.

Keywords—Logistics Services, Process Mining, Bottleneck, Competitiveness, Event Log

                                                                                                I.         Introduction

In recent decades, with the growth and expansion of information technology and the development of global and national markets, logistics services have become one of the main pillars of competition in many industries. Organizations and logistics companies are constantly seeking ways to improve their methods and processes to provide more effective services while minimizing operational costs and time. These challenges and needs have highlighted the necessity of managing and improving logistics processes more than ever before. Although numerous efforts have been made in the past to enhance the technical aspects of processes, many of these solutions have failed to fully address the major challenges of logistics and meet the complex needs of this field .

Process mining, as an innovative and efficient approach, has proven to be effective in analyzing, monitoring, and improving processes. This tool is especially useful when process complexities and related challenges hinder optimal organizational performance. The primary goal of this research is to enhance the productivity of logistics companies by identifying and resolving process-related issues and reducing bottlenecks. Through process mining, event logs can be analyzed, and by identifying bottlenecks and process deviations, solutions can be provided to improve operational time and reduce costs.

The global logistics and transportation market is growing at a rapid pace and is expected to reach $8040.55 billion by 2030 . This growth indicates an increasing need for the improvement and efficient management of logistics processes. However, many organizations still face challenges such as accurately identifying bottlenecks and aligning processes with designed models. Costs associated with “Last-Mile Delivery,” which sometimes account for up to 50% of total logistics costs, highlight the high sensitivity of this area to any changes . Surveys conducted by private companies, including PCA Predict and Zetes, indicate that missed deliveries and delays are key factors contributing to increased logistics costs, emphasizing the importance of improving logistics processes and better managing them.

Previous research shows that process improvement can lead to improvements in three main indicators: speed, cost, and service quality—indicators that play a fundamental role in maintaining organizational competitiveness. In this regard, process mining, as a new and efficient tool, has the ability to analyze, monitor, and improve processes effectively. This method, especially in identifying and correcting bottlenecks and inconsistencies in organizational processes, can significantly enhance organizational performance [5]. In the logistics sector, using process mining and related tools like Fuzzy Miner can help analyze and improve transportation and logistics processes, identifying and correcting existing inconsistencies and errors.

The objectives of this research are to identify and resolve process bottlenecks and inconsistencies in logistics services using the process mining approach. This research analyzes and reviews the current state of logistics processes and provides solutions to improve performance and increase the competitiveness of these services. The focus of this study is on using real data and event logs to identify and correct processes and improve resource allocation. Additionally, by creating support tools for managers and experts, it becomes possible to identify and resolve bottlenecks, ultimately leading to improved logistics service quality and increased process efficiency.

Considering the limitations of this research, certain areas related to logistics services and process mining have not been addressed. These areas include the impact of logistics processes on global political and economic changes, consumer behavior and its effect on process design, environmental impact assessments of improved processes, and the use of emerging technologies such as deep learning alongside process mining. This study primarily focuses on the operational and technical aspects of process improvement and the application of process mining in this field.

The research questions and hypotheses revolve around examining the extent to which logistics services can be improved using process mining and analyzing the correlation between bottlenecks and contextual information in logistics services. The main hypothesis of this research is that by utilizing process mining algorithms, logistics processes can be analyzed and improved with greater accuracy, ultimately leading to reduced operational time, lower costs, and increased process compliance.

                                                                                        II.        Literature Review

A.    Introduction and Connection to Enterprise Architecture

In recent decades, process mining has emerged as a novel and efficient approach for analyzing and improving organizational processes. This approach, particularly in the field of enterprise architecture—which focuses on optimizing business processes and increasing efficiency—plays a crucial and indispensable role. Enterprise architecture, by providing a comprehensive and cohesive framework for analyzing, designing, and implementing business processes, creates an ideal foundation for the effective use of process mining [6]. The review and analysis of the connections between enterprise architecture and process mining not only contribute to a deeper understanding of the role of this tool in improving organizational performance but also facilitate the identification of weaknesses and bottlenecks in processes. In this regard, process mining, by accurately identifying these bottlenecks, offers optimal solutions for improving and enhancing organizational productivity [7]. This is particularly important in the complex and dynamic logistics environment, where logistics companies must continuously and swiftly adapt their processes to environmental changes and evolving customer needs to stay competitive.

B.    Principles, Tools, and Technologies of Process Mining

Process mining, as one of the powerful tools in the field of business process analysis and improvement, utilizes various techniques to effectively identify issues and enhance organizational performance . This method, relying on pattern discovery and the analysis of process-related data, encompasses several stages, ranging from data extraction and exploration to process discovery, compliance checking, performance analysis, and ultimately, process prediction and improvement. Advanced tools such as ProM, Disco, and Celonis are recognized as key software in process mining, each offering unique features and capabilities to analyze complex data and deliver accurate results. These tools enable organizations to analyze their processes with greater precision using real data and event logs, allowing them to utilize these analyses to improve efficiency and effectiveness. A comparison of these tools is provided in Table I.

• A Review of Process Mining Tools .

C.   Context-aware process mining

Context-aware process mining adds additional contextual information to traditional process mining techniques to enhance the accuracy of analyses. Traditional process mining focuses on extracting and analyzing event logs to discover, monitor, and improve business processes. However, these methods often overlook contextual factors that can significantly impact processes. Context-aware process mining fills this gap by integrating data such as time, location, and frequency of events, related communications, tools, devices, and actors. This approach greatly aids in identifying more accurate and meaningful patterns and insights, particularly in complex and dynamic environments such as logistics and manufacturing [12].

D.   Logistics Services and Repetitive Processes

In the field of logistics services, “Last-Mile Delivery” is considered one of the most sensitive and complex stages of the supply chain, playing a decisive role in logistics costs and customer satisfaction . This stage faces numerous challenges such as traffic congestion, air pollution, and high costs due to failed deliveries, all of which can significantly impact the overall performance of logistics companies. To address these challenges and improve the efficiency of logistics processes, companies have sought innovative solutions that not only reduce costs but also enhance customer satisfaction. Various methods, such as home delivery and delivery to shared points, have been reviewed and evaluated in this regard . These methods are designed with a focus on cost reduction and productivity improvement and are considered efficient solutions to address the existing challenges in the logistics sector. Such solutions can not only optimize the final delivery process but also improve the customer experience and maintain the competitiveness of logistics companies in today’s dynamic and challenging markets.

E.    Bottlenecks and Non-Conformance in Logistics

Bottlenecks, as obstacles to the optimal performance of processes, can play a significant role in reducing efficiency and increasing operational costs for organizations. Identifying and accurately analyzing these bottlenecks allows organizations to improve their efficiency and effectiveness by optimizing processes and resources. In this context, process mining is introduced as an efficient tool that can be used to identify bottlenecks and assess their impact on organizational performance .
In addition to bottlenecks, one of the other major challenges in organizational processes is their non-conformance with designed models. These inconsistencies can lead to a decrease in service quality, an increase in costs, and ultimately a decline in customer satisfaction. Process mining, through a detailed analysis of real data and event logs, not only helps identify these inconsistencies but also offers effective solutions to align processes with organizational goals and improve their compliance. This closer alignment of processes with the designed models leads to enhanced service quality, reduced unnecessary costs, and overall improvement in organizational performance. The quality of a process model is typically described considering four dimensions:

• Fitness: The ability to observe the behavior of event logs in the discovered model.
• Precision: Preventing irrelevant behavior from being included in the event log used in the discovery process, such as avoiding the concept of underfitting.
• Generalization: The capacity to accept new similar events related to previous events used for discovery, like avoiding the concept of overfitting.
• Simplicity: Keeping the model as simple as possible.

Conformance checking can be used to align the model with the reality obtained from the event log in the system . According to reference, conformance verification can be employed to validate documented processes, identify process deviations, point to various instances, and attempt to find their common points. Additionally, conformance verification can be used to calculate the efficiency of a discovered process model and improve the new or existing model. Conformance checking is used in multiple cases, making it one of the pillars of process mining.