Real-Time Electrical Energy Optimization in E-Commerce Systems Based on IoT and Mobile Agents

Abstract

The integration of the Internet of Things (IoT) into mobile agent technology has fundamentally transformed the landscape of e-commerce by enabling intelligent, adaptive, and energy-efficient solutions. In this paper, we present a new system for integrating the information-sharing capability of IoT-enabled devices with the advanced abilities of mobile agents for the optimal utilization of energy when conducting e-commerce activity. The mobile agents are used as a mediating agent in the transaction and will capture operation data to share with stakeholders (not in the transaction) who might be able to provide services in association with that transaction. The operational data is collected, stored, and analyzed in real-time via IoT devices, facilitating adaptive decision-making while providing continuous monitoring of the system and servicing to improve energy management, efficiency, and operational performance. The combined IoT and energy capacity will enhance data sharing and provide more energy-efficient activities. The evaluation of the system was completed through simulations, as well as through real-world scenarios, achieving a decrease of approximately 27.8% in total energy consumption and savings of over 30% on operational costs. Moreover, the proposed architecture achieved a reduction of up to 38.9% for response times for resource management, under load, while also demonstrating a 50% reduction in response time for real-time event handling. Therefore, the effects of the proposed approach have been proven to be effective through simulations and real-world case studies, showing improvements in energy consumption and costs, as well as flexibility and adaptability. The findings of this study show that this framework not only minimizes energy consumption but also maximizes scalability, responsiveness to user demands, and robustness against variability in an e-commerce workload. This effort illustrates the potential for extending the lifetimes of e-commerce infrastructures and developing sustainable e-commerce models, demonstrating how IoT-based architectures can facilitate better resource allocation while achieving sustainability goals.

1. Introduction

The rapid advancement of the Internet of Things (IoT) has revolutionized industries worldwide, with e-commerce emerging as a primary beneficiary of these developments. As online retail platforms expand in scale and complexity, their operational demands— particularly in regards to energy consumption—have surged, posing challenges to both economic efficiency and environmental sustainability [1,2,3]. Addressing these issues necessitates innovative strategies such as energy optimization, adaptive decision-making, and real-time data processing in e-commerce infrastructures. In this context, autonomous and intelligent software entities, known as mobile agents, have emerged as transformative solutions for enhancing e-commerce operations [4]. These agents operate within IoT-enabled systems to dynamically optimize relevant processes, including data analysis, decision-making, and resource management. By integrating mobile agents for context-aware analysis with IoT-based real-time data collection, e-commerce platforms can transition toward sustainable and energy-efficient operations without compromising scalability or performance [5].

Energy consumption is a significant issue in e-commerce, as large infrastructures provide 24/7 access to meet consumer demand. This constant operational state leads to substantial electricity consumption, increasing operational costs and environmental risk. In light of the global shift towards carbon-neutral practices and sustainable technology, achieving energy efficiency in this digital business environment is prudent for both economic considerations and environmental sustainability. Intelligent technologies in the form of IoT and mobile agents will provide a pathway to mitigate both issues simultaneously.

Both mobile agents and the IoT are included in this study due to their complementary strengths, real-time data management, device distribution, and autonomous decision-making. The IoT allows for the seamless, sensor-based monitoring of the environment, and mobile agents offer autonomous intelligence and adaptability for responding to tasks. Used together, the IoT and mobile agents enable the decentralized, contextualized energy management of resources, allowing for the more dynamic management of energy, but they will still need to operate under a defined set of rules. This tendency towards individual instance versatility provides the IoT and mobile agents with a natural affinity for working together within distributed systems that display the ability for context changes. These systems must accommodate dynamic information processing, and they are well suited to meet the demands of energy-efficient, scalable omnichannel operations. The convergence of mobile agents and IoT technologies presents significant advantages. IoT devices provide continuous monitoring and data collection across e-commerce infrastructures, while mobile agents leverage this data to implement proactive adjustments, such as the optimization of energy consumption during peak and off-peak hours [6]. This decentralized approach reduces latency, enhances system reliability, and is well suited to the high demands of modern e-commerce. However, IoT-driven mobile agent systems also present critical challenges, particularly concerning data security, privacy, and system resilience. Given the growing threat of cyberattacks targeting IoT environments, it is imperative to develop robust frameworks that safeguard sensitive data and ensure operational integrity [7,8,9].

Beyond cost reduction, the drive for operational efficiency in e-commerce extends to environmental responsibility, especially in an industry characterized by intense competition and evolving consumer expectations [10]. Energy-intensive systems, such as large-scale server operations, logistics, and inventory management, have become key targets for optimization. By integrating the IoT and mobile agents, e-commerce platforms can monitor energy usage across various components, predict future consumption patterns, and implement intelligent waste-reduction strategies. This approach not only yields financial benefits but also aligns with sustainable business practices, which are increasingly vital for meeting global environmental regulations. Moreover, the adaptability and scalability of IoT-driven mobile agent systems enable their application across diverse e-commerce contexts [11]. These technologies offer specialized solutions for complex challenges, from managing warehouse energy consumption to optimizing transportation routes for fuel efficiency. This flexibility allows businesses of all sizes—from start-ups to large corporations—to leverage advanced technologies for competitive advantage [12].

Recent studies underscore the benefits of integrating agent-based modeling with the IoT to enhance e-commerce strategies. The use of mobile agents significantly improves adaptive decision-making, enabling businesses to dynamically optimize inventory management and enhance customer satisfaction [13]. The integration of IoT and mobile agents fosters the development of intelligent ecosystems in which resource efficiency and operational effectiveness are enhanced through data-driven methodologies. Concurrently, the research emphasizes the importance of efficient resource utilization in heterogeneous environments, such as IoT and cloud computing. Mobile agent-based frameworks dynamically manage resource consumption, helping e-commerce systems to reduce energy usage while improving overall infrastructure efficiency [14].

Furthermore, sensor-fusion techniques play a crucial role in enhancing IoT-enabled systems by facilitating precise energy management. The integration of sensor data strengthens decision-making capabilities, as exemplified by the HERACLES algorithm, which employs context-aware data fusion to optimize energy adaptation [15]. IoT technology is thus instrumental in advancing sustainable e-commerce solutions through sophisticated sensor-fusion and real-time energy management techniques.

Additionally, the convergence of artificial intelligence (AI) and blockchain technology is essential for ensuring secure and transparent data sharing within IoT ecosystems. The integration of AI-driven analytics with blockchain technology enhances data accuracy and reliability [16]. This combination ensures that mobile agent-based IoT networks adhere to stringent security protocols, protecting e-commerce operations from cyber threats and unauthorized access.

By adopting these advanced methodologies, the IoT and mobile agent systems have the potential to revolutionize energy efficiency in e-commerce, striking a balance between environmental sustainability and business competitiveness. The four-layer architecture of an IoT system, illustrated in Figure 1 [17,18,19], outlines the distinct roles performed by each layer within the IoT environment.

✓

Sensing Layer: This layer is devoted to data collecting and consists of embedded devices such as sensors and actuators. These gadgets collect environmental data that is essential to system functioning.

✓

Network Layer: This layer employs short- and medium-range communication systems such as Wi-Fi, Bluetooth, and Zigbee and integrates long-range communication systems such as LoRaWAN and cellular networks (4G/5G) for wide geographic coverage. Strong security measures include encryption and authentication protocols for the exchange of data.

✓

Data Processing Layer: Primarily cloud-based, this layer performs intensive data analysis to derive actionable insights. It supports advanced computational operations using data management systems, machine learning algorithms, and data lakes that store large volumes of raw data.

✓

Application Layer: In this layer, development could be carried out in an endless array of general-purpose programming environments, such as Python, Java, MATLAB, or other domain-specific tools, depending on the application requirements. Any information about specific versions has been omitted to accommodate improvements made across the various platforms.

This paper offers a novel framework that combines IoT and mobile agent techniques to handle energy optimization in e-commerce. By combining intelligent decision-making, real-time monitoring, and sophisticated security measures, the framework lowers energy use and complies with international sustainability standards. The research shows, through simulations and real-world implementations, how this framework may revolutionize e-commerce operations and open the door to innovative, safe, and sustainable business ecosystems in the age of the Internet of Things.

In this paper, we make several significant contributions to advancing sustainable e-commerce systems through the integration of Internet of Things (IoT) technologies and intelligent control mechanisms. Overall, our simulation results illustrate a reduction in total energy consumption of 27.78%, a decline in operational costs of 30.56%, and an enhanced real-time system responsiveness of 51.25%. First, we designed and implemented an IoT-based monitoring framework capable of collecting and analyzing real-time energy usage data across diverse e-commerce platforms. To facilitate proactive energy management, we developed autonomous mobile agents capable of modifying the operational states of energy-intensive subsystems such as HVAC and lighting systems, with simulation results indicating a reduction in energy consumption exceeding 27%. Additionally, we introduced adaptive mobile agents with generalizable design features to accommodate heterogeneous IoT devices and varying operational requirements, thereby supporting scalability from localized to large-scale deployments. A neural network-based machine learning model was also incorporated to forecast energy demand within the monitored systems, enabling predictive control strategies and optimized energy allocation. Furthermore, we implemented a secure communication protocol to ensure authenticated, encrypted, and synchronized interactions between IoT devices, agents, and e-commerce backends. To enhance user engagement, we developed personalized energy efficiency recommendations based on sensor-derived data, including user behavior and ambient environmental conditions, thereby increasing responsiveness to system settings. Finally, we demonstrated, through design case studies and simulation experiments, which our framework effectively reduces energy waste and supports internationally recognized environmental sustainability objectives.

The remainder of this paper is organized as follows. Section 2 reviews the historical context and relevant literature, with a particular focus on contemporary IoT-enabled solutions for energy optimization in e-commerce environments. Section 3 presents the proposed system architecture and algorithms, emphasizing key components such as mobile agents and their coordination with IoT devices. Section 4 outlines the experimental setup, reports the preliminary results, and describes the validation frameworks employed at multiple integration points. Section 5 discusses the current challenges and outlines future directions for enhancing and scaling IoT-driven e-commerce platforms. Finally, Section 6 concludes the paper by summarizing the main findings and suggesting avenues for future research.

2. Related Work

This section reviews the literature on IoT-driven e-commerce solutions, with a particular focus on the integration of mobile agents for energy optimization. It examines several relevant studies to identify gaps in the current research, especially in applying mobile agents and recommendation systems to improve energy efficiency in e-commerce environments. The review highlights the potential of these technologies to optimize energy consumption and enhance operational efficiency in IoT-based e-commerce systems.

2.1. Energy Optimization in IoT and E-Commerce Logistics

The study conducted in ref. [20] investigates the optimization of energy service distribution systems within e-commerce logistics by integrating theoretical modeling and practical experimentation. Specifically, it addresses system cost reduction of simultaneous delivery models through the strategic selection and configuration of delivery sites, with a particular emphasis on vehicle routing and site location optimization. The research employs a two-stage hybrid heuristic algorithm that combines genetic algorithms, maximum coverage techniques, and cluster analysis to achieve its objectives. Simulation results validate the efficiency of the model, highlighting its effectiveness in mitigating traffic congestion, reducing environmental impact, enhancing the efficiency of urban logistics, and addressing the challenges of last-mile delivery.

2.2. Machine Learning in Retail Logistics

The authors of [21] propose a comprehensive model for optimizing forward and reverse logistics networks in e-commerce, with a focus on achieving environmental, social, and economic sustainability. The model is formulated as a mixed-integer linear programming (MILP) problem aimed at minimizing energy consumption, carbon emissions, and operational costs associated with vehicle routing. Empirical evaluation using a hypothetical case study demonstrates that the model effectively balances all three sustainability dimensions with only a 9.02% deviation from the optimal potential for each objective. The findings underscore the model’s utility in supporting sustainable logistics operations by ensuring efficient vehicle deployment and optimal routing strategies.

2.3. Smart Distribution in Industrial IoT with Blockchain Technology

In [22], the authors address the dynamic vehicle routing problem inherent in logistics by introducing a time-windowed dynamic routing model. The approach partitions dynamic time frames into static periods, facilitating iterative optimization and energy reduction. Simulations using the Solomon dataset and Java-based tools show a 3.6% improvement in vehicle utilization and an 83.8 mile reduction in trip distance. The proposed method demonstrates robust performance in managing dynamic routing challenges, yielding significant energy savings and cost efficiency while enhancing system responsiveness.

2.4. Logistics Routing with Multi-Agent Systems

The research in [23] presents a cloud-based logistics scheduling model embedded with IoT capabilities. The model focuses on minimizing processing time and energy consumption while optimizing task scheduling and load balancing to improve resource utilization. A key innovation is the integration of circular chaotic mapping into the chameleon swarm optimization algorithm (C3SOA), which enhances the search process. Experimental results indicate that the proposed approach outperforms benchmark models in regards to load balancing and resource efficiency, making it well-suited for dynamic IoT-enabled logistics environments.

2.5. IoT Applications in Warehouse Logistics Management

Study [24] critiques the limitations of traditional warehousing systems in China and demonstrates how IoT integration can significantly enhance operational efficiency. The adoption of IoT technologies facilitates improved inventory tracking, space utilization, and real-time activity monitoring. These enhancements enable data-driven management, support predictive analytics, and contribute to the digital transformation of warehouse logistics. The study underscores the potential for IoT to create more agile and responsive supply chains within the evolving logistics sector.

2.6. Energy Efficiency in the Retail Sector Through Policy and Demand-Side Strategies

The research in [25] highlights the growing emphasis on energy sustainability within the retail sector, which historically has not been a major contributor to energy demand. The study discusses the role of energy audits and retrofitting policies, particularly the adoption of energy-efficient technologies such as LED lighting and HVAC upgrades. Regulated energy audits are shown to be effective in identifying energy-saving opportunities, while collaborative initiatives between government entities and retailers are promoted as mechanisms for supporting long-term sustainability goals and reducing operational costs in retail energy use.

2.7. Demand-Side Energy Management Models in Smart Retail Cities

The authors of [26] introduce a demand-side management (DSM) aggregator model designed for small- and medium-sized retail enterprises whose energy usage is often underreported. The model enables the use of centralized energy management programs across multiple retail locations and leverages deep learning techniques to generate precise energy demand forecasts, achieving a mean squared prediction error below 2.05%. The proposed solution offers scalability for energy service companies (ESCOs) operating in smart cities, facilitating load balancing and enhancing both economic and environmental performance across the retail sector.

2.8. Energy Optimization in Supermarkets Using IoT-Based REMS

In [27], the authors present a real-time energy management system (REMS) based on IoT architecture, targeting supermarkets with high energy demands stemming from HVAC and refrigeration systems. Through the application of big data analytics and AI-driven control mechanisms, the system achieves substantial energy savings—up to 25% for air handling units and 33% for water pumps—with an overall energy consumption reduction of 17%. The REMS also offers a financially viable solution with a three-year payback period and demonstrates scalability for broader retail applications.

2.9. Smart Grid Energy Management Through AI, IoT, and Blockchain Technology

Study [28] proposes an integrated smart grid framework that synergistically combines artificial intelligence (AI) for predictive analytics, the Internet of Things (IoT) for real-time data acquisition, and blockchain technology for secure and transparent energy transactions. The model leverages advanced machine learning algorithms—specifically long short-term memory (LSTM) networks, extreme gradient boosting (XGBoost), and support vector machines (SVM)—to enhance demand forecasting and manage solar energy through reinforcement learning mechanisms within photovoltaic microgrids. This integrated, multi-technology solution achieves a 21% increase in energy accessibility and a 28% reduction in peak load, underscoring its potential for improving energy efficiency and system stability in smart city applications.

To ensure a valid and equitable comparative analysis, only studies that report quantitative results specifically related to energy consumption and latency performance have been included in

Table 1. Comparative analysis of selected studies reporting quantitative metrics regarding energy and latency benchmarked against performance targets of the proposed solution.

Authors	Technique Used	Adaptation	Latency Improvement (%)	Mobile Agent	Energy Savings (%)	Performance	Scalability
[20]	IoT-based sensor network for real-time tracking.	√	×	√	×	√	×
[21]	Machine learning with IoT integration for predictive analytics.	√	√	×	√	√	×
[22]	Mobile agent system for dynamic task allocation.	×	√	√	×	√	×
[23]	Hybrid IoT and mobile agent framework for e-commerce optimization.	×	√	√	√	×	√
[24]	IoT-based REMS for energy management in supermarkets.	√	×	√	×	√	×
[25]	Policy-based DSM in retail, with energy audits and conservation (e.g., HVAC, LEDs).	√	×	×	15%	×	√
[26]	DSM aggregator model with deep learning for forecasting in smart cities.	√	×	×	×	√	√
[27]	RFID-based IoT system for product inventory tracking.	√	×	×	17%	√	√
[28]	Multi-agent AI-IoT-blockchain smart grid architecture.	√	×	×	21%
Proposed Approach	The approach uses mobile agents in IoT-driven e-commerce to optimize energy use through data analysis, task management, and decision making.	√	27.78%	√	38.89%	√	√

. This criterion ensures alignment with the performance-focused objectives of the proposed system. The selected studies are compared against the present approach to illustrate how each addresses analogous challenges in mobile agent-based, IoT-driven e-commerce systems, particularly in the domain of energy management. Where available, numerical indicators of improvement—such as percentage reductions in energy consumption and latency—are provided. A check mark (√) indicates that a study addresses a particular performance challenge, while a cross (×) denotes its absence.

3. Materials and Methods

In this paper, we propose a control framework for IoT-driven e-commerce solutions including mobile agents, focusing on energy optimization. The system integrates nonlinear dynamic modeling with machine learning to enhance energy management. The IoT devices and mobile agents are modeled using nonlinear differential equations. The hybrid control strategy combines traditional feedback mechanisms with a neural network that dynamically adjusts energy optimization parameters based on real-time sensor data. The neural network is trained on historical energy consumption patterns and IoT device performance, enabling it to predict and implement adjustments for improved energy efficiency in e-commerce operations.

3.1. Algorithm 1: Optimized Energy Management in IoT-Driven E-Commerce Using Mobile Agents

Algorithm 1 employs mobile agents to optimize energy management in IoT-driven e-commerce platforms. To ensure low energy consumption, the main goal is to effectively assign jobs to devices, depending on their energy levels and proximity to the task site. While devices with adequate energy are chosen to complete tasks, others are monitored for low energy levels and refilled as needed. To promote coordinated activities inside the e-commerce system, the algorithm also allows devices to communicate their status and task outcomes. This method enhances the system’s overall sustainability and efficiency by optimizing energy consumption, which makes it appropriate for real-time e-commerce systems using mobile agents.

To improve transparency and reproducibility, we present Algorithm 1 in a Python-style pseudo code format. This semi-structured format provides an explicit overview of the logic of the mobile agents in assessing the energy levels of the device and assigning proximity-based tasks based on thresholds, as well as managing the tasks to recharge the devices. While Algorithm 1 represents a simpler version of the proposed energy optimization strategy, providing a useful guide for implementation, the essential logic remains representative.

Algorithm 1. Optimized Energy Management in IoT-Driven E-commerce

import time def initialize_system(devices):       for device in devices:             device.update({‘status’: ‘active’, ‘energy_level’: 100, ‘task_history’: []}) def select_optimal_device(devices, task):       eligible = [d for d in devices if d[‘energy_level’] > 20 and d[‘proximity’] == task[‘location’]]       return max(eligible, key=lambda d: (d[‘energy_level’], -len(d[‘task_history’]))) if eligible else None def allocate_tasks(devices, tasks):       for task in tasks:    if (device:= select_optimal_device(devices, task)):      device[‘energy_level’] -= task[‘energy_required’]      device[‘task_history’].append(task[‘id’])      print(f”Task {task[‘id’]} assigned to Device {device[‘id’]}.”)    else:      print(f”[WARNING] No device found for Task {task[‘id’]}.”) def report_status(devices, server):       server[‘log’] = [{‘id’: d[‘id’], ‘energy’: d[‘energy_level’], ‘tasks’: d[‘task_history’]} for d in devices]       print(f”Status updated for {len(devices)} devices.”) def recharge_devices(devices):       for d in devices:    if d[‘energy_level’] < 25:      time.sleep(1)      d[‘energy_level’] = 100      print(f”Device {d[‘id’]} recharged.”) def main():       devices = [{‘id’: i+1, ‘proximity’: p, ‘energy_level’: e} for i, (p, e) in enumerate([(‘A’, 80), (‘B’, 50), (‘A’, 90)])]       tasks = [{‘id’: f’T{i+1}’, ‘energy_required’: r, ‘location’: l} for i, (r, l) in enumerate([(30, ‘A’), (40, ‘B’), (50, ‘A’)])]       server = {}       initialize_system(devices)       allocate_tasks(devices, tasks)       report_status(devices, server)       recharge_devices(devices)       print(“[FINAL LOG]:”, server[‘log’]) main()

3.2. Architecture of Proposed System

The proposed system architecture for optimizing energy consumption in IoT-driven e-commerce solutions integrates several key components designed to monitor, process, and analyze energy usage in real-time. The architecture efficiently collects data from IoT devices embedded within the system using mobile agents, enabling dynamic energy optimization. The system combines real-time data processing with energy management algorithms to minimize consumption while maximizing operational efficiency. The system ensures adaptive energy optimization across various e-commerce operations by leveraging mobile agents that dynamically adjust control parameters. The architecture is designed to be scalable, effective, and capable of providing actionable insights for energy management, as illustrated in Figure 2.

Table 2. Overview of the components of the proposed system.

N	Component	Description
1	E-commerce platform	The platform that manages products, customers, transactions, and recommendations, allowing personalized customer interactions and integrating IoT data for energy optimization.
2	IoT devices (sensors and IoT nodes)	Devices that collect environmental data (temperature, humidity, energy consumption) and IoT nodes for local data processing and aggregation.
3	Mobile agents	Autonomous software entities that analyze data and make decisions to optimize energy consumption in the e-commerce environment.
4	Communication mechanism	Mechanism for secure data transmission between IoT devices, the e-commerce platform, and mobile agents, ensuring communication between all components.
5	Control systems	Systems that implement actions prescribed by mobile agents to optimize energy consumption, such as turning off equipment or adjusting temperature.
6	Energy monitoring tools	Tools that visualize and analyze energy consumption, providing reports and visualizations to identify energy optimization opportunities.

provides a comprehensive overview of the primary parts of our proposed system. Each part plays a role in maintaining the efficiency, scalability, and real-time performance of the system.

3.3. Types of Agents

The proposed system combines agent-based modeling with the Internet of Things for energy optimization, enhancing the usefulness and effectiveness of e-commerce platforms.

Figure 3 provides a general overview of the communication structure of an Internet of Things (IoT)-driven e-commerce system, in which several agents work together to improve different parts of the system. Every agent—represented as a robot—has a specific function, including gathering data, managing energy, monitoring the system, making decisions, and ensuring security. The arrows show the information flow and job distribution between various agents, showing how they cooperate to accomplish system goals, including secure communication, energy efficiency, and real-time decision-making. This cooperative structure allows the system to work efficiently by coordinating each agent’s skills.

Table 3. Roles of agents in IoT-driven e-commerce system for energy optimization.

Agent Name	Primary Role
Data Acquisition Agent (DAA)	Collects data from IoT devices integrated into the system.
Energy Management Agent (EMA)	Analyzes energy data to optimize consumption.
Security Agent (SA)	Ensures the protection of sensitive data and secure communications.
Device Control Agent (DCA)	Manages IoT devices by adjusting their parameters or states.
Monitoring Agent (MA)	Monitors system performance and detects anomalies.
Communication Agent (CA)	Manages communication between agents and external systems.
Simulation Agent (SimA)	Conducts simulations to predict the impacts of energy management strategies.
Logistics Optimization Agent (LOA)	Optimizes energy-intensive logistics processes such as transportation and warehousing.
Costumer Interface Agent (CIA)	Provides a costumer interface for system monitoring and configuration input.
Decision-Making Agent (DMA)	Coordinates strategic decisions to ensure alignment with global goals.

contains detailed information about each agent in the system, including their name and primary role. It acts as a reference for understanding the purpose and functionality of each agent within the system architecture.

3.4. Flow Chart of the Proposed Model

Figure 4 presents a flowchart of the comprehensive architecture for the approach, focusing on applications for energy optimization. It systematically displays the interconnected layers and processes, from data acquisition through IoT devices to actionable recommendations and control. The flow includes data collection, secure transmission via the network layer, and data validation. Key components such as data processing, employing advanced machine learning and cloud platforms, and decision-making through mobile agents are highlighted. The architecture also integrates error handling to ensure robust operation in cases of invalid data. This flowchart effectively demonstrates the systematic approach to optimizing energy usage in e-commerce environments using IoT technologies.

4. Results

This section presents the experimental results, their interpretation, and our conclusions. It focuses on the performance of IoT-driven e-commerce solutions with mobile agents for energy optimization. The findings demonstrate the effectiveness of nonlinear models and neural network-based optimization in improving energy efficiency. The conclusions highlight the potential benefits of scalable e-commerce operations and suggest future energy management enhancements using IoT technologies.

4.1. Performance Evaluation of IoT-Driven E-Commerce Solutions

This section thoroughly assesses the performance of IoT-driven e-commerce solutions for improving energy consumption. Critical criteria will be evaluated, including cost reduction, operational efficiency gains, and energy savings. Data collected from IoT sensors integrated into the system will be examined to demonstrate how real-time data might influence energy management tactics.

4.1.1. Experimental Setup and Parameters

The experiments were conducted in a custom Python-based simulation environment that mimics a real-time, IoT-based e-commerce system. Its general framework employs a number of mobile agents, each with a specific set of tasks, as identified in Table 3. The agents interact in two operational zones (A and B) to perform various tasks for collecting data, analyzing energy consumption, and activating devices. Task execution followed the criteria requiring proximity to task location and a minimum energy threshold of 20%. The devices were recharged prior to circumstantial task assignments when the energy threshold fell below 25%. Simulated tasks included inventory adjustments, environmental monitoring, and logistics streamlining, with each task requiring energy ranging between 30–50 W/h. The simulations were run continuously in controlled time intervals to assess varying operational loads and events such as demand increases. The metrics recorded included the consumption of energy, the efficiency of the assignment tasks and criteria, agent decision latencies, and the real-time task fulfilment response service time.

The research outlined in the above empirical example utilized the simulated real-time behaviors and configurations described in

Table 4. Experimental setup and simulation parameters.

Parameter	Description
Simulation Environment	Custom Python-based simulation framework.
Agent Framework	JADE (Java agent development framework) for multi-agent system implementation.
Number of Agents	Five agents: DAA, EMA, DCA, MA, and DMA.
Agent Functions	Data collection, energy analysis, device control, monitoring, decision-making.
Operational Zones	Two zones (A and B), simulating e-commerce warehouse or logistics areas.
Number of Devices	Three IoT devices with unique IDs and location tags (e.g., Zone A or B).
Energy Requirement per Task	A range of 30–50 W/h, depending on task type (inventory, environment, logistics).
Energy Threshold for Tasking	A ≥20% energy required to accept new tasks.
Recharge Trigger Level	The device is recharged if energy drops below 25%.
Simulation Duration	A 48 h per test scenario, with 1 min time steps.
Performance Metrics	Energy consumption (W/h), task success rate, agent response latency, real-time adaptability.

. These parameters represent the key simulation parameters and key test configurations and were in place to assess the behavior and performance of each of the five JADE-based agents in a real-time IoT based e-commerce scenario.

4.1.2. Selection Criteria and Parameter Justification

To assure the practical relevance, functionality, and repetition of the experimental approach, we prepared representative test and simulation parameters that reflect the operational conditions of a modern e-commerce system. The study investigates three scenarios representing energy-aware task allocation, multi-agent cooperation in multiple zones, and responsiveness to sudden changes in load. These scenarios are consistent with the scenarios encountered in smart warehouses and logistics networks.

The energy thresholds chosen comprised a 20% relative to load capacity for eligible tasks and a 25% load for recharging tasks. The amount of energy required for these tasks ranges between 30–50 W/h, which is reflective of pre-existing consumption characteristics for lighting, HVAC controls, and logistics tasks.

The two zones (A and B) are representative of areas within a warehouse environment. Therefore, the agents exhibited the ability to make decisions while taking energy and proximity into consideration. The simulation was run for 48 h, with 1 min reporting intervals, providing a sense of the responsiveness of the agent over a short time frame, and over the 48 h period, actionable optimization of the system would occur. Therefore, the test was both realistic and capable of being scaled and repeated.

4.1.3. Energy Efficiency Improvements and Optimization

This section aims to measure the energy effectiveness enhancement resulting from realizing IoT-driven solutions. Using information from IoT sensors and smart devices, it will examine energy usage patterns before and after the optimization techniques are implemented. The effectiveness of these solutions in lowering energy usage throughout the e-commerce platform will be evaluated by examining key indicators such as peak load management, energy consumption per transaction or operational job, and reduction in energy waste.

The energy usage of different parts of the suggested IoT-driven framework and the conventional e-commerce system is contrasted in

Table 5. Energy consumption comparison between traditional methods and the proposed framework.

Component	Energy Consumption (Traditional)	Energy Consumption (Proposed Framework)	Energy Savings (%)
IoT Devices	120 W/h	90 W/h	25%
Mobile Agents	150 W/h	100 W/h	33.33%
Data Transmission	80 W/h	60 W/h	25%
Processing Units	100 W/h	75 W/h	25%
Total Energy Usage	450 W/h	325 W/h	27.78%

. It draws attention to the energy savings made possible by the use of mobile agents, IoT devices, and effective data processing and transmission techniques.

Figure 5 presents a comparative analysis of energy consumption between a conventional system and the proposed IoT- and mobile agent-based e-commerce architecture. The data are visualized using a smoothed (tension) curve. The y-axis represents energy usage in watts per hour (W/h), while the x-axis corresponds to key system components (e.g., IoT devices, mobile agents, etc.). The smoothed curves clearly depict the energy savings achieved by the proposed framework, highlighting its superior energy efficiency relative to that of the conventional system.

The baseline energy consumption data for the traditional system are derived from prior studies [21,23,24], which were based on static architectures lacking mobile agents and adaptive control mechanisms. In contrast, the proposed architecture demonstrates substantial reductions in energy usage across multiple components. These improvements are attributed to the dynamic coordination and intelligent control enabled by mobile agents operating within the IoT ecosystem.

4.1.4. Operational Cost Reduction

This section evaluates the expanded operating impacts, which mainly include cost savings and enhanced resource exploitation. Mobile agents can dynamically optimize operational choices, leading to cost reductions by integrating real-time data collecting and analysis from IoT sensors. This section will assess how these systems enhance resource allocation and overall system performance while reducing operating costs connected to labor, energy, and resource waste. The analysis will demonstrate the relationship between improved cost efficiency and optimal energy use.

The operating costs of the suggested IoT-driven framework and the conventional system are contrasted in

Table 6. Operational cost comparison between traditional methods and the proposed framework.

Cost Category	Traditional System (USD)	Proposed Framework (USD)	Cost Savings (%)
Data Collection	500	350	30%
Energy Costs	800	500	37.5%
Maintenance and Support	300	250	16.67%
System Administration	200	150	25%
Total Operational Cost	1800	1250	30.56%

. It illustrates the overall savings by breaking down expenses into categories such as data collecting, energy use, maintenance, and administration.

The suggested framework and the conventional system’s operating cost reduction curves are smoothed out in Figure 6. Included on the x-axis are expense categories (such as “Data Collection”, “Energy Costs”, etc.), and on the y-axis, the costs expressed in dollars are displayed. The smoothed curves show the cost savings resulting from the suggested framework when compared to the costs connected to typical systems, especially when it comes to data collecting, energy, and maintenance.

The operational cost values for the conventional system are derived from the IoT-based e-commerce models reported in [21,23,24], all of which lack mobile agent optimization and decentralized control mechanisms. These studies are used as reference benchmarks for each cost parameter, enabling a consistent and comparative evaluation of the proposed system’s economic impact. By contrasting these baseline values with those obtained through decentralized, intelligent task identification, and coordination in the proposed architecture, the analysis demonstrates the potential cost efficiencies enabled by the integration of mobile agents within IoT-driven e-commerce systems.

4.1.5. System Scalability

This section addresses and appraises the adjustability of the IoT-driven e-commerce solution as the business increases. It will consider the system’s ability to manage an increasing number of devices, transactions, and energy needs. The solution’s long-term effects on general efficiency and sustainability will also be investigated.

Table 7. System response time under different loads in the traditional vs. proposed framework.

Load Scenario	Response Time (Traditional)	Response Time (Proposed Framework)	Latency Reduction (%)
Small-Scale (10 Devices)	500 ms	350 ms	30%
Medium-Scale (100 Devices)	800 ms	550 ms	31.25%
Large-Scale (1000 Devices)	1200 ms	750 ms	37.5%
Extra-Large Scale (5000 Devices)	1800 ms	1100 ms	38.89%
Average Response Time	1150 ms	725 ms	37.39%

contrasts the reaction times of the suggested framework with those of conventional systems for a range of load scenarios, from small-scale to extra-large-scale deployments. The outcomes show that even with higher system loads, the framework can retain minimal latency.

Figure 7 shows the scalability of the suggested and conventional systems, depending on the load situations (e.g., “Medium-Scale,” “Small-Scale,” etc.). The y-axis displays the system’s reaction time (in milliseconds), while the x-axis displays various load conditions (number of connected devices). The suggested framework outperforms the conventional system, as seen by the smoothed curves, which show improved scalability and reduced reaction times as the number of devices rises.

Baseline latency values for the traditional system for several levels of processing loads were taken from the reports in [21,23,24], as limitations regarding central coordination and task processing create bottlenecks as the system scales.

4.1.6. Real-Time Responsiveness

This component assesses the system’s responsiveness to dynamic events, including changes in inventory, demand spikes, and the reallocation of energy resources.

The reaction times for several real-time events in the suggested framework and the conventional system are displayed in

Table 8. Real-time event response time comparison between the traditional and proposed framework.

Event Type	Response Time (Traditional)	Response Time (Proposed Framework)	Improvement (%)
Customer Demand Surge	1200 ms	650 ms	45.83%
Inventory Update	900 ms	450 ms	50%
Energy Resource Reallocation	1000 ms	500 ms	50%
System Failure Recovery	1500 ms	750 ms	50%
Average Response Time	1200 ms	587.5 ms	51.25%

. It shows how the system can swiftly adjust to dynamic changes such as spikes in consumer demand and reallocations of energy resources.

Figure 8 shows the real-time responsiveness of both the proposed and traditional systems for all event types, such as inventory updates, high customer demand, energy resource reallocation, and failure recovery. The y-axis indicates the system response in milliseconds (ms), while the x-axis highlights the event categories. The smoothed curves convey that the proposed framework is more responsive to any dynamic operational change, outperforming the traditional system across all events. Specifically, the proposed system shows a reduction in response time from 45.83% to 50% for the events category, with an overall average improvement of over 51.25%. These results validated the framework’s improved ability to respond to real-time events more quickly and efficiently, which are the most important aspects of energy awareness and performance sensitivity in e-commerce.

The traditional system using time events to represent the reaction time for real-time events is based on the latency measures provided in [21,23,24], where agent-based event processing is not included. The improvements in our system are apparent in regards to both speed and adaptability.

4.1.7. Generalizability of Results Across Operational Scenarios

The operational scenarios illustrated in Table 5 through Table 8 are designed to generalize the system’s performance across representative e-commerce environments. These scenarios encompass core operational dimensions such as energy consumption, operating costs, system scalability, and real-time responsiveness—each of which is critical to ensuring the sustainability and efficiency of modern digital commerce ecosystems.

The evaluated components reflect typical elements found in contemporary e-commerce infrastructures: Internet of Things (IoT) devices, cloud-based data transactions, and dynamic system responsiveness. These features are foundational to the current and future footprints of smart logistics and e-commerce platforms. Importantly, the scalability tests simulate realistic growth trajectories, spanning small-, medium-, and large-scale deployment environments. This variation enables a more robust understanding of how the system performs under different levels of operational intensity.

In addition, the system’s responsiveness to event-driven scenarios—such as sudden surges in demand or the need for rapid recovery—demonstrates its applicability to real-world operational disruptions. These systemic stressors are increasingly relevant in the context of volatile e-commerce demand patterns and logistical bottlenecks.

By integrating a diverse set of operationally meaningful and variable scenarios, the analysis aims to move beyond narrow case-specific findings and toward broader generalizability. This ensures that the proposed system’s efficacy is relevant not only within controlled experiments but also across a range of practical deployment contexts relevant to e-commerce and smart logistics.

As depicted in Figure 9, which compares traditional frameworks with the proposed IoT-based architecture in terms of energy consumption, operational costs, scalability, and real-time responsiveness, the benefits of the new system are evident. The proposed solution achieves up to a 33% reduction in energy consumption, a 30% reduction in operational costs, and a 51% improvement in real-time responsiveness. These improvements substantiate the architecture’s capability to support more adaptive, efficient, and scalable e-commerce operations across diverse scenarios.

Figure 9 summarizes the benchmark performance of the traditional system, based on average values reported in [21,23,24]. These studies serve as consistent baseline references across all evaluated dimensions, enabling a structured and equitable comparison. The results highlight the advantages of integrating a mobile agent-based component into the proposed architecture, revealing notable improvements in system performance and operational efficiency compared to the results of conventional models.

4.2. Mobile Agent’s Creation and Integration

4.2.1. Creating Mobile Agents

The section shows that the creating of mobile agents involves the development of software that can autonomously move between systems to perform various tasks and collect data. Security and communication mechanisms are essential considerations in their design, as shown in Figure 10.

4.2.2. Interactions Between Agents

Usually, when a mobile search agent and a location agent communicate, the mobile search agent requests location information from the location agent. After receiving this request, the location agent returns the required location information to the mobile search agent. When the mobile search agent receives this data, it can be used to improve its search functions. Both agents can work together efficiently through this connection when location-based information is needed, including the use of geolocated search and navigation, as shown in Figure 11.

5. Discussion

The integration of IoT-enabled e-commerce systems with energy optimization strategies, particularly those leveraging mobile agents, offers substantial operational efficiencies and reductions in energy consumption. Our experimental findings demonstrate a 27.78% improvement in energy savings and a 38.89% enhancement in system responsiveness, confirming the viability of mobile agent systems, as previously supported by Elhoseny et al. [21] and Hamza et al. [23]. These findings are especially significant in the context of small-scale e-commerce environments, where computational resources and operational burdens are more manageable.

However, scalability emerges as a critical issue when extending these solutions to larger and more complex e-commerce infrastructures. Despite improvements in computational capabilities and energy optimization algorithms, the system must effectively manage vast streams of real-time data and coordinate large numbers of mobile agents. This places high demands on both system architecture and user expertise, particularly as operational complexity increases.

In addition to performance, system vulnerabilities and energy trade-offs warrant attention. IoT infrastructures are inherently susceptible to sensor failures, connectivity disruptions, and other malfunctions, which may compromise decision-making regarding energy optimization. Moreover, IoT devices themselves consume energy continuously through sensing, data collection, and communication, which can offset some of the energy savings unless carefully managed.

Security is another significant concern. IoT-enabled mobile agent systems are vulnerable to various cyber threats, including DDoS attacks, sensor spoofing, and agent interception. Ensuring secure communication and protecting user privacy requires strong encryption, robust authentication protocols, and continuous monitoring. Future implementations will incorporate distributed authentication schemes and lightweight intrusion detection mechanisms tailored for mobile agent environments to address these risks.

Our findings align with the observations of Mahesh and Anilkumar [29], who emphasized the need for logistics systems to remain adaptable amid dynamic consumer demand, seasonal fluctuations, and evolving market conditions. To maintain robust performance under such variability, IoT-enabled mobile agents must employ flexible control algorithms capable of learning and adapting to changing operational contexts.

Although the long-term benefit-to-cost ratio of the proposed system is favorable, the initial investment in IoT infrastructure remains a barrier—particularly for small and medium-sized enterprises (SMEs). Lowering the cost and simplifying deployment through modular system design and standardized components will be critical to enable broader adoption.

6. Limitations of the Proposed System

Although pilot studies of IoT-based e-commerce solutions using mobile agents to optimize energy consumption are encouraging, considerable limitations remain.

6.1. Technical and Operational Constraints

Despite the promising outcomes demonstrated in pilot studies, the proposed IoT-based energy optimization system using mobile agents faces several technical and operational limitations that must be addressed before large-scale adoption.

Scalability remains the foremost limitation. While the system performs effectively in small- to medium-scale environments, its feasibility in large-enterprise-level deployments—characterized by complex workflows and high transaction volumes—has not been established. Validation under real-world conditions involving high data loads is essential to confirm the system’s robustness, coordination efficiency, and ability to operate under stress.

The reliability of the IoT infrastructure presents a second major challenge. Sensor errors, data inconsistencies, or communication failures can compromise the accuracy of energy optimization decisions. Additionally, while IoT devices contribute to system-level efficiency, they also introduce their own energy demands, which must be factored into the overall energy balance. Cybersecurity and data privacy risks further complicate deployment; strong encryption, secure communication protocols, and rigorous authentication procedures are critical for safeguarding the data exchanged between agents and devices.

Real-time data processing also poses a significant constraint. The high volume of sensor-generated data can create processing bottlenecks, delaying or impairing the timeliness of energy management decisions. This underscores the need for more efficient data processing mechanisms and infrastructure capable of supporting high-throughput, low-latency operations.

Adaptability to dynamic conditions, such as fluctuations in energy demand or consumer behavior, also requires improvement. Unexpected workload changes or usage spikes can reduce the system’s responsiveness. Future research should prioritize enhancing the flexibility of mobile agents, incorporating predictive algorithms, and developing adaptive strategies capable of responding swiftly to evolving operational environments. Lightweight agent models and hybrid cloud-edge architectures may offer a promising pathway by minimizing computational loads while maintaining responsiveness.

Deployment costs represent another significant concern. The initial investment required to implement the proposed architecture—especially compared to those of more conventional IoT-only systems—may act as a barrier to adoption, particularly for small and medium-sized enterprises (SMEs). To address this, future efforts should explore modular and scalable deployment models that allow incremental adoption and cost-effective implementation.

To enhance system availability, resilience, and scalability, several advanced technologies will be incorporated in future iterations. Edge computing will be utilized to offload real-time processing closer to the data sources, reducing both latency and reliance on central servers. The integration of 5G connectivity will enable high-speed, low-latency communication between IoT devices and mobile agents, thereby improving decision-making capabilities in dynamic environments. Moreover, the use of federated machine learning will allow agents to learn from decentralized data sources while preserving user privacy and data security. Together, these innovations—combined with lightweight agent designs and modular system architectures—will provide a robust foundation for developing a scalable, secure, and adaptive energy optimization framework.

6.2. Absence of Real-World Implementation

Although the proposed system has been thoroughly validated through simulations and controlled experiments, it has not yet been deployed in a real-world e-commerce or logistics environment. This limitation stems primarily from the absence of partnerships with industry stakeholders and restricted access to operational infrastructure and live data. As a result, key real-world factors—such as integration with legacy systems, live user interaction, and disruptions at the hardware level—remain untested. These considerations are vital for assessing the system’s practical effectiveness, robustness, and maturity within its intended operational context.

We acknowledge that live testing represents an indispensable step toward full system validation and realization. Without field implementation, it is difficult to fully evaluate aspects such as usability, system reliability, and cost-effectiveness under dynamic, real-world conditions. Accordingly, future work will prioritize collaboration with logistics service providers or e-commerce platforms to deploy the system in an operational setting. Such deployment will enable the collection of empirical evidence regarding system performance, identify integration challenges, and inform refinements for broader scalability and adoption.

7. Conclusions and Broader Implications

This study has examined the design and potential of an IoT-based energy optimization framework for e-commerce environments, utilizing mobile agents to manage and reduce energy consumption. By autonomously processing real-time sensor data, mobile agents dynamically adapt energy usage patterns across various operational components, contributing to more efficient and sustainable platform management. Experimental validation indicates that the integration of mobile agents into IoT ecosystems can substantially reduce energy-related costs while supporting environmental sustainability objectives.

Beyond the immediate technical achievements, the proposed architecture represents a shift toward decentralized, autonomous systems capable of operating in complex, data-intensive environments. This offers a meaningful departure from conventional centralized energy management approaches, suggesting new possibilities for real-time, adaptive decision-making in e-commerce logistics and platform operations. The system’s embedded recommendation capabilities further enhance its responsiveness to environmental and behavioral changes, positioning it as a promising model for intelligent energy governance.

While technical and implementation-related limitations have been discussed extensively in the preceding section, it is important to underscore the broader implications of this work. Specifically, the study contributes to a growing body of literature that emphasizes the role of distributed intelligence in managing critical resources in dynamic digital ecosystems. It also highlights the potential for mobile agent systems to be extended beyond energy optimization to other domains in which decentralized control, autonomy, and scalability are essential.

Future research could further explore the interdisciplinary dimensions of agent-based IoT systems, particularly the integration of behavioral, organizational, and regulatory perspectives. For instance, understanding how end-users and operational managers interact with such systems in live settings may inform more intuitive interfaces and adaptive learning models. Similarly, the evolving legal and ethical landscape surrounding autonomous decision-making in commercial environments merits closer attention, especially regarding algorithmic transparency and data governance.

In summary, while practical deployment remains a challenge, this research offers a conceptual and technical foundation for developing more autonomous, scalable, and sustainable e-commerce infrastructures. The findings open up important avenues for further inquiry across the technical, organizational, and societal dimensions of digital platform management.