Geographic Information System-Based Framework for Sustainable Small and Medium-Sized Enterprise Logistics Operations

Abstract

Dispatching goods is becoming more difficult to manage in the field of logistics due to the high demand for order shipments. This is related to the increasing popularity of the use of e-commerce platforms by consumers, where products are required to be delivered rather than being bought in physical stores. Dispatch management is one of the critical components in a supply chain since it covers the coordination of tasks among stakeholders from the warehouse to the consumer’s doorstep. In this study, the authors propose a framework leveraging geographic information to sustain logistics operations, specifically in terms of managing last-mile delivery and return trip orders. This includes scheduling, communications, and the inventorying of the shipment status of goods. A mobile application built on this framework was integrated with a waypoint order optimization algorithm considering an entire route that traverses all the required pick-up and delivery points. It was pilot tested with an actual dispatch operation of a logistics company, yielding decreases of 92% and 43% in the average turnaround time and carbon footprint per completed service request, respectively, a decrease of 57% in operations cost, and an increase of 72% in profit. With the adoption of this framework, this study aims to contribute to the overall efficiency and sustainability of logistics operations in a wider geographic range.

1. Introduction

Consumers of the present generation have grown more dependent on the use of e-commerce, given the convenience it offers in terms of the purchase and delivery of goods [1,2]. Consequently, the demand for logistics has exponentially increased since e-commerce businesses heavily rely on this industry [3]. The companies running e-commerce services have opted to make use of third-party logistics services instead of buying their own delivery vehicles and managing their own order shipments. Third-party logistics or 3PL is a type of logistics service that specializes in moving goods from source to destination points [4]. Its services commonly include freight forwarding, cargo delivery services, and point-to-point delivery [5]. Outsourcing this service has enabled e-commerce businesses to focus more on improving their core operations.

The process of bringing the product to the customer, technically known as last-mile delivery, has garnered a lot of attention recently because of its pivotal role in customer satisfaction, environmental impact, and cost-effective operations [6,7]. It is composed of several steps, as follows: (1) the storing of order information in the digitized central system, (2) the physical arrival of orders at the logistics company, (3) the assignment of orders to delivery personnel, (4) the updating of order statuses, and (5) the closing of orders for successful end-to-end transactions [8]. A return trip is added as an additional step when the product delivered is wrong or defective. Return trips are the transfer of products or assets from the vehicle’s destinations back to their origins or to another specified point [9,10]. Although it appears to be a simple step in the logistics process, return trips can bring about challenges and opportunities that may have a significant impact both on the business’ cost savings and on the environment. Coordination among stakeholders throughout the process is seen as crucial, and among the difficult tasks of a dispatcher are scheduling and fleet tracking [11,12]. Inefficiency in handling these tasks has contributed to operational overhead costs and to greater environmental carbon footprints [13].

According to one study, 25% of the total CO₂ emissions of an urban city are generated from the logistics network. This is because of the urban deliveries made using smaller trucks catering to individual households, as compared to traditional bulk and long-distance deliveries [14,15]. The more frequent the stops and idle times of the vehicles, the higher the fuel consumption and the higher the greenhouse gas contribution. This scenario has been evident in last-mile deliveries or the final leg of shipments [16,17,18]. Other contributing factors to the increased carbon footprint are the vehicle type, traffic characteristics, fuel type, and the fleet behavior [19]. There has long been a need to structure fleet dispatch and space usage at a single time of delivery to lower the contribution of logistics operations to pollution [20,21]. With the use of the IoT, the coordination between customers and drivers was made more real-time in one study [22]. Another study focused on managing received requests to make dispatch assignments more economical at the pick-up vehicle end [23,24]. In terms of routing optimizations, several search heuristics [25,26,27] have been proposed, not only in the transportation sector but also in supply distribution [28,29] and in fields that require the redirection of entities in general [30,31,32]. Dijkstra’s method is one of the most classic yet stable search algorithms that serve as the primary backbone of route-finding algorithms for Google Maps. It aims to find the shortest path from one point to another by visiting all the adjacent vertices and calculating the path cost at the same time [33]. The path cost may vary depending on the required metric. Most commonly, either distance or time cost metrics are used [34]. A review of related studies and lacking approaches is summarized in

Table 1. Review of related studies—summary.

Title	Discussion	Lacking Approaches
Dynamic Dispatch Algorithm Proposal for Last-Mile Delivery Vehicle [35]	Introduces an algorithmic solution to vehicle allocation, mathematical modeling, smart cities, and big data. Resonates with creating a robust and efficient last-mile delivery system and promotes the importance of technology-driven solutions that tackle societal, environmental sustainability, and economic issues.	The study lacks an actual implementation and relied on datasets with limited constraints in constructing the model. In addition, some factors for optimizing last-mile delivery were not considered such as traffic congestion and fuel usage.
Studying the Rerouting of Empty Carriers during their Return Trips to Manage Rare Mobile Resources in a Physical Internet [36]	Primarily focuses on addressing the challenge of efficiently moving carriers to where they can also be of use in a logistics network and utilizing the return trips of carriers, which are often underutilized and may travel empty or partially empty. Mirrors the innovative approach of leveraging return trips of carriers to optimize resource distribution within the logistics network. It reflects the researchers’ aim of efficient resource allocation in last-mile delivery.	The study needs follow-up research to design an actual experiment to prove the feasibility and effectiveness of the proposed solutions to optimize logistics resource distribution.
From City Logistics to a Merger of Smart Cities and Smart Logistics from City Logistics to a Merger of Smart Cities and Smart Logistics [37]	This study highlights the emergence of sustainable logistics management and the exploration of the components essential in creating a balance between economic growth and preservation of environmental and security aspects of logistics. Addresses consumer needs such as on-demand deliveries and their desire for speed and flexibility in an omnichannel environment. In addition, it takes note of the growing consumer perceptions of sustainability and global warming.	Only studies the components required to create a sustainable logistics management system and has no implementation of the solutions proposed.
Optimal Transfer Point Locations in Two-Stage Distribution Systems [38]	The research paper explores how to further utilize unmanned aerial systems in last-mile retail distribution networks. Strengthens the approach of optimizing resource allocation within last-mile delivery operations, offering valuable insights for software development and emphasizing the economic and efficiency benefits of delivery optimization.	The study does not apply to ground vehicles, which are in need of solutions. Hence, it cannot be adopted in existing logistics operations.
Minimizing Last-Mile Delivery Cost and Vehicle Usage Through an Optimized Delivery Network Considering Customer-Preferred Time Windows [39]	Introduced a cluster-based approach to optimize last-mile delivery while simultaneously considering time windows and other practical requirements such as package compatibility.	The study needs a follow-up with an actual implementation, as it used limited datasets in simulating the model. Important factors in optimizing the last-mile delivery such as traffic density and fuel consumption were not considered.
Smart City—Smart Logistics Amalgamation [40]	Stated that smart logistics has a positive influence on the performance of intelligent cities when integrated. The proposed solution is within the scope of smart logistics since it is a software application that benefits the stakeholders in the logistics industry and may also influence smart cities positively.	The study focused on Melbourne and considered only business retailers, suppliers, and service providers as the stakeholders.
Logistics Distribution Route Optimization Method for Peach Products Transport [41]	Utilizes the ant colony algorithm to optimize the route for peach product delivery. Route optimization is one of the ways to reduce operating costs, which is in line with the application.	The study needs a follow-up with an actual implementation since it relies only on datasets with limited constraints to simulate the model. The algorithm cannot be tested in real-time.
Preparing for the smart cities: IoT enabled last mile delivery [42]	This study suggests a conceptual framework that combines Global Positioning System (GPS), Internet of Things (IoT), and Transportation Management System (TMS) technologies to address the lack of descriptive delivery addresses in many locations and the reliance on antiquated, semi-digital methods for tracking and customer interactions by offering real-time visibility, dependability, and intelligent last-mile solutions.	The report recognizes the limitations of the data that were gathered for the investigation.
Multimodal Autonomous Last-Mile Delivery System Design and Application [43]	This study investigates creative last-mile food delivery methods employing autonomous cars (robot-based, drone-based, and hybrid delivery). Eighteen scenarios with various fleet sizes and demands were examined in an agent-based simulation using Mississauga as a case study. Includes last-mile delivery optimization and a comparison between delivery models.	The study refers to stringent laws governing the use of drones for commercial purposes, which pose a significant restriction. It is crucial to continue investigating legal frameworks and potential modifications or adaptations to allow for autonomous food distribution systems.
Unmanned Aerial Vehicle Last-Mile Delivery Considering Backhauls [44]	This study proposes the FSTSP-B (Backhaul Deliveries in the Flying Sidekick Salesman Traveling Problem) concept, which considers both delivery and pick-up requirements for unmanned aerial vehicles (UAVs) in logistics. Presents a new concept that incorporates backhaul deliveries.	The adjacent distance-based assignment (ADA) heuristic can manage instances with up to 100 customer nodes. However, it is essential to further improve the algorithm’s effectiveness and scalability to cope with greater datasets for larger-scale logistical operations.
No-Load Backhaul Ratio Reduction by Interactive Routing for Small Goods Delivery with Window Time Constraints [45]	This study introduces a novel interactive routing model that uses a movable storage house to dynamically match delivery trucks, lowering backhauling costs and increasing efficiency under time window restrictions.	The paper does not mention the interactive routing model’s scalability. It is crucial to ascertain how effectively the suggested strategy can manage more extensive and more complicated logistics networks as delivery volumes increase.
A last-mile delivery problem with alternative delivery options based on prospect theory [46]	The goal is to reduce overall costs, including transportation expenses and fines for customer complaints, and to increase the competitiveness of logistics companies. To solve the model effectively, the paper offers an improved adaptive extensive neighborhood search method (I-ALNS). Results demonstrate that I-ALNS boosts solutions well.	The study did not evaluate the suggested method’s scalability to address the challenges of more extensive logistics networks with high orders.
Route optimization for last-mile distribution of rural E-commerce logistics based on ant colony optimization [47]	This paper talks about finding a feasible route for last-mile distribution using ant colony optimization. Based on the results, the ant colony optimization reported increased profits for the logistics enterprise.	This study lacks an actual implementation of the proposed solution.
Route optimization using Saving Matrix Method—A Case study at Public Logistics Company in India [48]	The Saving Matrix Method allowed the company to save on the overall distance and monthly travel expenses. The algorithm used in this study significantly benefited the companies in India.	The study does not include traffic jams and assumes that they do not affect the speed of the truck.
Digitalization and third-party logistics performance: exploring the roles of customer collaboration and government support [49]	Digitizing third-party logistics positively affected the financial and service performance in China.	Future research needs to specify and discuss thoroughly the type of digitization the company performed.
Impact of Using Third Party Logistics Provider on Organizational Performance in Bulgarian the Context [50]	The study proves through thorough statistical analysis that small and medium-sized manufacturing businesses perform better when outsourcing their supply chain integration, inventory management, transportation, and warehousing.	The research mainly focuses on small and medium-sized manufacturing companies in Bulgaria. The study’s conclusions are less broadly applicable to other sectors, firm sizes, or geographical regions.
Innovative solutions to increase last-mile delivery efficiency in B2C e commerce: a literature review [51]	Provides multiple solutions in different categories involved in last-mile delivery. The study focuses on providing variables for computing the overall transportation costs.	The study should be extended with more suggestions with regard to algorithms for faster and more efficient routing for last-mile delivery.
Integrated planning of inbound and outbound logistics with a Rich Vehicle Routing Problem with backhauls [52]	Tested multiple strategies to solve vehicle problems with backhauling. The best solution provided a 2.7% decrease in costs. The study used integrated inbound and outbound logistics planning for a company with backhauling.	The study lacks software and an actual implementation of the solution.
A multiple objective transportation problem approach to dynamic truck dispatching in surface mines [53]	Uses a dispatch management system for dump trucks that utilizes databases, a fleet system, and operational data management.	The study lacks an actual implementation of the fleet management system. The study’s system might have a different effect in real-world scenarios.
Applicability of Digital Tracking System on Third Party Logistics (TPL) Services [54]	In this study, the usage of software applications for tracking delivery operations was evaluated. It was found that digitization of the process can help reduce waste such as time and other resources.	The study focuses on tracking the fleet and does not include actual management of trip requests.
Blockchain Adoption Intention by a Third-Party Logistics Company: A Malaysian Case Study [55]	The research is another digitization approach towards 3PL with blockchain as its reference. It offers new ways to track deliveries, such as with Facebook and WhatsApp to receive updates.	The study focuses on security and does not solve the efficiency problem in logistics operations.

.

The strengths and weaknesses of the existing studies relevant to the authors’ proposed solution are presented in this chapter. After thorough research, we gained an understanding of ways of dealing with the issues in dispatch management and gained insights into how various algorithms and methods can be integrated into solutions mitigating the problems encountered in logistics operations in general.

This study aimed to address the pressing economic and environmental needs and challenges within the logistics industry, primarily for small and medium-sized third-party logistics businesses, through a geographic-information-system-based framework [56]. A software application was developed and integrated with a waypoint optimization algorithm to manage the workload in last-mile delivery and return trips [57,58] more efficiently, leveraging actual geospatial data such as the location and traffic-based travel time, which can be obtained from this framework. Concurrently, we promote green logistics with a reduction in carbon emissions from courier activities on the road while maximizing the logistics resources at hand [59]. The achievement of these objectives is evident with the resulting metrics presented in Section 3 of this paper.

2. Materials and Methods

A logistics company was interviewed to validate the pain points identified in the previous chapter in the context of dispatch management and to gain insights into the proposed framework, which will be discussed further in Section 2.2. A dispatch management software application was developed using Android Studio (Hedgehog v2023.1.1) as an Integrated Development Environment (IDE), Flutter (v3.19.0) as the UI Software Development Kit (SDK), and Dart (v3.3) as the primary programming language, and this was powered by Google APIs (application programming interfaces) such as Maps, Places, Directions, and Firebase Cloud Messaging. The final prototype was tested, and its performance results were compared with the traditional approach of the identified logistics company in dispatch management.

2.1. Current Method Used by the Partner and Other Third-Party Logistics Companies

The traditional return trip and last-mile delivery transactions of most logistics companies like 5RJS Lanuza Logistics Corp. have been made through calls, Short Message Service (SMS), and emails from customers. The 3PL is directly contacted by the customer via phone, when they make a call to schedule the delivery of their goods. The dispatcher from the 3PL then arranges a vehicle and driver to send the products in a batch to the respective end customer’s location. The car usually returns to its origin point with empty or unused space after all assigned orders are successfully delivered. In the case of return trips, the dispatcher is contacted again by the customer, requesting a new schedule to transport goods from the destination back to the vehicle’s origin. The current method has presented limitations and challenges in establishing effective communication, coordination, and scheduling, resulting in inefficiencies and missed opportunities, especially in terms of optimizing delivery routes and use of carrier space. Additionally, transactions made via phone calls, SMS, and emails are difficult to track, and further issues in managing logistics processes are posed.

Similar conventional methods are being used in other small to medium-sized 3PL business operations in the country, such as F2, Majestic Group Global, Ernest, and SuperHawk Logistics. There were attempts to use software solutions for cargo-booking and delivery-tracking visibility for the end user. However, the dispatch management internal to the operations remains conventional. On the driver’s side, the ordering of deliveries is decided based on the expert judgment of the dispatcher, which is not reliable at all times, given the sporadic changes encountered on the road daily. Other 3PL software solutions, such as Lalamove (v112.3.0) and Grab Express (v5.332.0) services, utilize GIS to optimize delivery routes, although trip requests are received by drivers on an on-demand basis, meaning that the current transaction needs to be completed first before receiving another job, since dispatch management per se is not included in the software solution functions.

The old framework shown in Figure 1 has yielded a high turnaround time and low concurrent order processing. Using phone calls has limited single customer transactions as these calls take time, usually ranging from a few minutes to an hour in duration, depending on the concerns of the customer. The following describes the setup of the corporation:

The customer can only make inquiries about delivery orders through phone calls or SMS to the company. High call waiting times have become a common problem when there are simultaneous requests from multiple customers. When an inbound call comes through, a customer service representative (CSR) is assigned to interact with the customer and maintain good relationships. They answer customer concerns, relay messages to and from the management, provide service information, and monitor the capacity of trucks to handle a given number of deliveries. All these are crucial in maintaining a good business image for customers. Next, dispatchers interact with drivers while assigning delivery requests to available drivers, monitor the progress of deliveries, and relay the routes/paths that the drivers should take. Fleet tracking is performed using a whiteboard only. This is inconsistent and difficult to sustain, especially when the business is expanding. On the driver’s side, a call is made by the dispatcher when an accepted delivery order is placed, whether it is for last-mile or return trip delivery.

2.2. Proposed GIS-Based Framework

Four roles (see

Table 2. Roles and capabilities in the proposed application software.

Roles	User Access
Admin/Dispatcher	Accept or decline customer’s delivery requests. Assign the accepted/accommodated delivery request to the drivers. View customer’s pending delivery requests, available drivers, and progress of deliveries. View, update, block/unblock, and delete user accounts in the application software.
Driver	Create an account. Accept or decline trips assigned to them. View pending tasks. View routes and trip history. Update availability status and progress of delivery.
Customer	Create an account. Submit delivery requests. View the progress of the delivery requests.

) have been identified at the software level: administrator, dispatcher, customer, and driver. The customer is the one who requests the trips and monitors the progress of the delivery requests. The dispatcher will either accept or decline the delivery requests, depending on the availability of vehicles and drivers. The dispatcher will also be responsible for receiving and assigning requests to the available drivers, which they will then monitor. The driver will accept or decline the delivery requests depending on their condition and update their progress while delivering the goods. Lastly, the administrator has the same responsibilities as the dispatcher but has additional ones, specifically retrieving, updating, and deleting user accounts in the application software.

The proposed improvement to the framework of the 5RJSL Lanuza Logistics Corp. is shown in Figure 2. This utilizes mobile applications and GIS-driven algorithms at the backend to improve the business response times and increase the number of concurrent customers the business can handle. With the new framework, customer can send orders through the application software, and acceptance notifications are received without repeated dialing. The option to use SMS, calls, or emails for inquiries about concerns and special deliveries is still available. The customer service representative may answer special concerns through SMS, calls, and emails. The dispatcher can handle multiple orders at the same time as long as delivery trucks are available. The dispatcher is also responsible for assigning driver trips, including return trips through the application software, managing the truck fleet using the dashboard, and monitoring the trucks and customer orders. Lastly, the truck driver can update the dispatcher regarding delivery and availability much more quickly instead of calling the dispatcher back. They can also accept/reject customer orders, which is useful in case the driver needs to go home early or has health problems that arise.

To optimize the trip of a driver, the application software suggests a route plan based on various input parameters such as vehicle type, shipment locations, defined constraints, and vehicle cost parameters. The constraints include driver working hours, vehicle capacity, and window hours. This optimization is achieved through the waypoint order optimization algorithm, wherein a vehicle is assigned cost per hour, per kilometer, and per traveled hour parameters. The pseudocode representation of the algorithm is as follows:

• request.truck [0].startLocation -> transitions [0] -> visits [0] ->
• transitions [1] -> visits [1] -> transitions [2] -> … -> visits [n] ->
• transitions [n] -> request.truck [0].endLocation.

The startLocation and endLocation represent a truck’s origin and return point, respectively, after all orders have been shipped to the corresponding end customer location. Based on the GPS locations of pick-up/delivery points, the waypoint order is arranged to achieve the minimum travel hour and distance costs. These costs are extracted using Google’s Distance Matrix and Directions API. A transition is tagged between the truck’s current location and the following location to visit. Upon completion of a pick-up or delivery, it is tagged as a visit. This represents the assigned work that a truck must complete at the requested place and time. Google’s interactive polyline encoder utility was used to aid the driver in visualizing the path per visit.

Utilizing the GIS-based framework can sustain logistics operations with its combined workload management tool and shipment route optimization to increase time efficiency, cost savings, customer/driver satisfaction, and environmental friendliness through reduced carbon emissions despite being in the transportation sector.

2.3. Roles and User Access in the Application Software

The roles and capabilities of the application software are described in Table 2 and illustrated in Figure 3. The admin has the ability of a dispatcher plus the ability to view, update, block/unblock, and delete the details of the users in the database.

2.4. Dataflow in the GIS Framework

Figure 4 shows the data flow diagram of the application software. This figure illustrates the interaction between the database, APIs, different variables, and user data. The data start with the creation of users and drivers in the application. At first, the user will input their details such as email address, username, phone number, and password, and the driver will input the details mentioned earlier in addition to car details such as plate number, model, and their picture so that the admin and dispatcher can further validate their legitimacy and identify them more easily in the system once recorded in the database. The Mapping and Database module houses the algorithms and core GIS, which are the key contribution and difference made by the proposed study as compared to existing 3PL solutions of SMEs, whether conventional or with the aid of software applications.

2.5. Developed Application Interface

The log-in pages of the mobile application for the user and driver have been developed with a similar design but with separate backend forms for credentials’ validation. Once successfully logged in, the user and driver are redirected to a map, as illustrated in Figure 5. On the users’ side, they can see the available drivers within their area, indicated by a car icon. On the other hand, there is a “Go Online Now” button for the drivers. The driver can click it to indicate their availability to pick up goods from customers. The search destination page in Figure 6 is where a user can enter their preferred pick-up and destination points. The search algorithm predicts the location based on the characters being entered and dynamically lists places that match for easier and more accurate entry of locations, as shown in the pseudocode below.

FUNCTION searchLocation (locationName)         IF the length of locationName is greater than 1 THEN             CONSTRUCT the API URL for Google Places Autocomplete                 using the locationName and the Google Maps API key;             SEND a request to the Places API using the constructed URL;             AWAIT the response from the Places API             END IF             IF an error occurs during the API request THEN                 RETURN to the function             END IF             IF the response status from the Places API is “OK” THEN                 EXTRACT the predictions from the response;                 MAP each prediction to a PredictionModel object;                 CONVERT the mapped predictions to a list;                 UPDATE the dropOffPredictionsPlacesList state variable                     with the list of predictions             END IF END FUNCTION

A web application has been developed to manage trip requests on the admin or dispatcher’s side. The panel dashboard is shown in Figure 7. Here, the trip details such as trip ID, username, pick-up and drop-off addresses, and request status can be seen and managed by the admin to facilitate delivery service. Once a trip request is approved, the pseudocode below is initiated to find and notify the nearest online driver from the place of pick-up. A time-out is incorporated to limit the waiting time and immediately pass on the service request to the next-closest available driver. The notification dialog that pops up when a trip has been assigned to a driver is shown in Figure 8. When the driver accepts the request, the trip details container showing information on the driver’s estimated time of arrival, name, and car is populated on the user app, as shown in Figure 9. It also has a call button to enable communication between the user and the driver. When the user clicks the button, they are redirected to the phone’s dialing app, and the mobile app automatically pulls up the driver’s number. When the dispatcher rejects the trip or none of the drivers accept the request within the set time limit, a notification dialog will pop up stating that no driver is available. Figure 10 shows the navigation map of the driver when multiple requests were accepted within the given period. The routes highlighted are based on the resulting order of pick-up and delivery points using the waypoint order optimization algorithm, as discussed in Section 2.2 of this paper.

FUNCTION initializeGeoFireListener ()         INITIALIZE GeoFire with the database reference “onlineDrivers”;         QUERY nearby drivers using the current latitude and longitude of the user with a radius of 22 units; LISTEN for changes in the nearby drivers query results IF the driver event is not null THEN     RETRIEVE the child event type from the driver event;     SWITCH CASE on the child event type         IF the event is Geofire.onKeyEntere THEN             CREATE a new OnlineNearbyDrivers object;             POPULATE the object with the driver’s uid,                 latitude, and longitude;             ADD the driver to the nearbyOnlineDriversList;             IF nearbyOnlineDriversKeysLoaded is true THEN                 UPDATE the drivers on the Google Map         END IF         IF the event is Geofire.onKeyExited THEN             REMOVE the driver from the nearbyOnlineDriversList;             UPDATE the available online drivers on the map         END IF         IF the event is Geofire.onKeyMoved THEN             CREATE a new OnlineNearbyDrivers object;             UPDATE the object with the driver’s UID,                 latitude and longitude;             UPDATE the location of the driver in the nearbyOnlineDriversList;                     UPDATE the available online drivers on the map                 END IF                 IF the event is Geofire.onGeoQueryReady THEN                     SET nearbyOnlineDriversKeysLoaded to true                     UPDATE the available online drivers on the map                 END IF             END SWITCH CASE         END IF END FUNCTION

A proof-of-delivery function was also added to allow drivers to capture images and send them accordingly, along with the drop-off time and location. This is crucial when assessing disputes about lost packages or return requests for wrong/damaged packages.

2.6. Data Collection Methods

We replicated an average trip of a 4-wheeler vehicle of the partner logistics company. The scenario is that there are seven accepted service requests in a single day identified with a reference number, three of which have exact pick-up locations but with different delivery areas and lead times, and one is a return request, as shown in

Table 3. Test scenario for the combined last-mile and return trip requests.

Reference No.	Leadtime	Pick-up	Delivery Area
11224-1B	0 to 1 day	Br. Andrew Gonzales Hall	within the Metro
11224-2B	0 to 1 day	University Pad Residences Taft	within the Metro
11224-3B	0 to 1 day	Vista GL Taft by Vista Residences	within the Metro
11324-1B	1 to 2 days	Br. Andrew Gonzales Hall	outside the Metro
11324-2B	1 to 2 days	University Pad Residences Taft	outside the Metro
11324-3B	1 to 2 days	Vista GL Taft by Vista Residences	outside the Metro
11324-4B	1 to 2 days	Robinsons Place Manila	*Return Request

. All packages that cannot be delivered within the day, including those for return, will be dropped off at Mega Pacific Freight Logistics, Inc. (Manila, Philippines). Based on the requests received, the dispatcher has arranged the pick-up and delivery points in the form of a job request for the next-day schedule of an assigned driver. This will be further discussed in Section 3.1.

Data gathering is divided into three parts. In the first part, the researchers followed the order of job requests based on the expert judgment or familiarity of the dispatcher with the service locations without using the developed software or any navigation apps. In the traditional approach, the driver should be following this, but there are instances when the points are reordered depending on several factors such as the driver’s spatial awareness, traffic conditions, and inability to contact the recipient on their number. In the second part, the researchers used the software application, which receives the service request, allows the dispatcher to accept requests, assigns the requests to a driver, and organizes the route for the driver. The elapsed time, distance traveled, and fuel expenses were recorded, and the corresponding carbon footprint, service turnaround time, job requests completed, elapsed time savings, and projected profits were measured. It should be noted that due to time constraints, the researchers started outside the Br. Andrew Gonzales Hall in the actual testing for both the traditional and the proposed GIS-based frameworks, instead of Mega Pacific Logistics, Inc. The deviation in time and distance projected by Google Maps were added to the final results. In the last part of the data gathering, the researchers demonstrated the software application to several employees of the partner logistics company. They distributed a survey about user acceptance to them.

2.7. Freight Rate Matrix Provided by the Partner Logistics Company

The freight rate matrix provided by the partner logistics company is summarized in

Table 4. Freight rate matrix without toll fees for 4-wheeled vehicles provided by the partner logistics company.

Distance in km (2-Way)	Fuel Expenses (PHP)	Fuel Expenses (w/Reclaimed Tax) (PHP)	Labor Cost (PHP)	Total Expenses (PHP)	Freight Rate (PHP)	Freight Rate (Inc. of VAT) (PHP)	Final Freight Rate (PHP)	% Profit per Transaction
5	40.63	36.16	955.00	991.16	1338.06	1498.63	1500.00	33%
10	81.25	72.31	955.00	1027.31	1386.87	1553.30	1600.00	37%
15	121.88	108.47	955.00	1063.47	1435.68	1607.96	1700.00	41%
20	162.50	144.63	955.00	1099.63	1484.49	1662.63	1700.00	36%
25	203.13	180.78	955.00	1135.78	1533.30	1717.30	1800.00	39%
30	243.75	216.94	955.00	1171.94	1582.12	1771.97	1800.00	35%
35	284.38	253.09	955.00	1208.09	1630.93	1826.64	1900.00	38%
40	325.00	289.25	955.00	1244.25	1679.74	1881.31	1900.00	34%
45	365.63	325.41	955.00	1280.41	1728.55	1935.97	2000.00	37%
50	406.25	361.56	955.00	1316.56	1777.36	1990.64	2000.00	34%

. The researchers used this to generate the projected profits for testing using the current dispatch method. Another factor measured was the fuel cost per trip. To compute the fuel cost incurred in the traditional method, the researchers followed the fuel efficiency values and fuel price set by the partner logistics company. According to R.J. Lanuza in a personal interview on 19 February 2024, the company currently has the fuel efficiency set to 8 km/L and the fuel price set to PHP 65.00/L in computing the fuel expenses, assuming no applicable toll fees. The third column in Table 3 is the fuel expense less the reclaimable amount equal to the consumption tax of around 11% for businesses. The labor cost is calculated as the sum of the driver’s and helper’s fees of PHP 550.00 and PHP 405.00, respectively. These are the current fixed amounts being paid by the partner logistics company in exchange for the services of the assigned driver and helper, regardless of the total distance traveled. The freight rate that is being charged to customers is computed as the sum of total expenses and the marked-up amount. The mark-up pricing of the partner logistics company is set to 35% of the total costs. The freight rate is also subjected to 12% VAT when the customer uses the service. The final freight rate charged to the customers per trip is simply the freight rate, including VAT, rounded up to the nearest hundred. In the final column, the percentage profit per transaction is computed as the final freight rate less the 12% VAT and divided by the total expenses. It can be seen that the profit per transaction ranges from 33 to 41% in the conventional method.

2.8. Freight Rate Matrix Provided by the Researchers

Table 5. Freight rate matrix without toll fees for 4-wheeled vehicles proposed by the researchers.

Distance Travelled	Fuel Expenses (PHP)	Fuel Expenses (w/Reclaimed Tax) (PHP)	Driver’s Fee (PHP)	Helper’s Fee (PHP)	Maintenance Fee (PHP)	Total Expenses (PHP)	Base Freight (PHP)	Per km Rate of PHP 55.00	Total Freight Rate (PHP)	Total Freight Rate (Inc. of VAT) (PHP)	Final Freight Rate (PHP)	% Profit per Transaction
5	40.63	36.16	160.00	105.00	30.00	331.16	200.00	275.00	475.00	532.00	600.00	59%
10	81.25	72.31	190.00	120.00	45.00	427.31	200.00	550.00	750.00	840.00	900.00	85%
15	121.88	108.47	220.00	135.00	60.00	523.47	200.00	825.00	1025.00	1148.00	1200.00	102%
20	162.50	144.63	250.00	150.00	75.00	619.63	200.00	1100.00	1300.00	1456.00	1500.00	113%
25	203.13	180.78	280.00	165.00	90.00	715.78	200.00	1375.00	1575.00	1764.00	1800.00	121%
30	243.75	216.94	310.00	180.00	105.00	811.94	200.00	1650.00	1850.00	2072.00	2100.00	128%
35	284.38	253.09	340.00	195.00	120.00	908.09	200.00	1925.00	2125.00	2380.00	2400.00	133%
40	325.00	289.25	370.00	210.00	135.00	1004.25	200.00	2200.00	2400.00	2688.00	2700.00	137%
45	365.63	325.41	400.00	225.00	150.00	1100.41	200.00	2475.00	2675.00	2996.00	3000.00	140%
50	406.25	361.56	440.00	245.00	170.00	1216.56	200.00	2750.00	2950.00	3304.00	3400.00	146%

shows the freight rate matrix with the use of the developed software application for dispatch management. This was also used to generate the projected profits, to test the dispatch performance using the proposed GIS-based framework. The existing freight charges of the partner logistics company are based on a two-way or return trip (origin to destination and vice versa), making them high despite the fact that deliveries are made from one pick-up/delivery point to another rather than from the origin (often from the warehouse) to the corresponding delivery locations and vice versa. In addition, the fixed costs paid to the driver and helper per trip without reference to the actual distance traveled significantly contribute to the freight charges despite the fact that some packages require minimum effort from the helper and some delivery points are very near each other, which does not necessarily equate to a roundtrip distance per se. More established or large logistics companies implement a base fare on every trip and add a fee for every kilometer traveled to the destination [60]. This makes it fair not only for the customers but also for the company. With the software application, it is easier to adjust the rates being paid to the drivers and helpers, as well as the rates being charged to the customers, since the actual data on the distance traveled between points can now be extracted.

The fuel efficiency rate of 8 km/L and price of PHP 65.00/L in calculating the fuel consumption still apply, as per the computation in (1).

Fuel Consumption Cost = Fuel Price × No. of Kilometers/Fuel Efficiency

(1)

With the use of a software application, a new variable cost was introduced: the maintenance cost set at 5% of the final freight rate. The fees paid for the driver and helper were converted from a fixed to a variable cost set at 10% with a base pay of PHP 100, and 5% with a base pay of PHP 75, respectively. The base pay and percentage value ratio were maintained at 4:3, as in the fixed labor fees of the driver and helper in Section 2.7. These were also set at par with the regular service rates of existing 3PL software solutions for every individual end-user booking request and in consultation with R.J. Lanuza in a personal interview on 7 March 2024. The total expense value was now computed as the sum of the labor and maintenance fees and fuel expenses with reclaimed tax. A base fee with additional charge per km distance pricing strategy was applied for the new freight rate matrix. A base fee of PHP 200 and an additional PHP 55 per km were set at par with the regular pricing of existing online package delivery systems for 4-wheeled vehicles such as Grab and Lalamove. The total freight was still subjected to 12% VAT and rounded up to the nearest hundred. With the new costing of labor and pricing of the freight service, the percentage profit was expected to be from 2 up to 5 times higher than the percentage profit when using the conventional method depending on the distance traveled while maintaining a reasonable freight charge per km distance traveled by the courier.

3. Results and Discussions

3.1. Elapse Time and Job Request

Job request numbers were mapped based on the order of pick-up/delivery services that must be completed within the day as scheduled by the dispatcher (old framework) or the software app (new framework). There could be more than one reference number mapped in a single job request. This means that there could be multiple customers requesting a service in a similar location. But only one driver and one helper could be assigned for the series of job requests in a single trip/route, as in job request numbers 0 to 6 in

Table 6. Summary of elapsed time for testing without the application software.

Job Request No.	Service Location	Service (P/D)	Elapsed Time (mins)	Reference No.
Pick-Up Location	Mega Pacific Freight Logistics, Inc.	-	0
0	Br. Andrew Gonzales Hall	P	40	11324-1B
1	University Pad Residences Taft	P	17.35	11324-2B
2	Vista GL Taft by Vista Residences	P	42.83	11324-3B
3	Robinsons Place Manila	P	18	11324-4B
4	Adamson University Main Building	D	14.35	11224-1B
5	SM Mall of Asia	D	47.97	11224-3B
6	SM City Bicutan	D	61.82	11224-2B
Drop-off Location	Mega Pacific Freight Logistics, Inc.	-	49.75	-
		Job Total Time:	292.07

and

Table 7. Summary of elapsed time for testing with the application software.

Job Request No.	Pick-up/Delivery Location No.	Service (P/D)	Elapsed Time (mins)	Reference No.
Pick-Up Location	Mega Pacific Freight Logistics, Inc.	-	0
0	Br. Andrew Gonzales Hall	P	35	11324/11224-1B
1	University Pad Residences Taft	P	8.23	11324/11224-2B
2	Vista GL Taft by Vista Residences	P	16.78	11324/11224-3B
3	Robinsons Place Manila	P	8.6	11324-4B
4	Adamson University Main Building	D	13.93	11224-1B
5	SM Mall of Asia	D	31.63	11224-3B
6	SM City Bicutan	D	52.85	11224-2B
Drop-off Location	Mega Pacific Freight Logistics, Inc.	-	19.75	-
		Job Total Time:	186.77

. The service column indicates whether the job to be completed is a pick-up or delivery of a package. The elapsed time column shows the recorded duration in minutes when a job is completed from the previous to the current point. When comparing Table 6 and Table 7, it can be observed that the time it takes to arrive at all of the service locations is significantly shorter with the use of the software application. The difference in the total elapsed time between testing with and without the app is 105.3 min, which corresponds to a 36.05% saving in time, while the job requests completed see an increase of 42.86%, as shown in

Table 8. Job request completed by framework.

	Turnaround Time
	Elapsed Time	Job Request
Old Framework	292.07	7
New Framework	186.77	10
% Difference:	36.05%	42.86%

. The average percentage of turnaround time savings per completed customer service request is 91.85%, as shown in

Table 9. Turnaround time per completed service request.

	Turnaround Time
Reference No.	Old Framework	New Framework	% Time Savings
11224-1B	852.53	47.54	94.42%
11224-2B	962.32	123.79	87.14%
11224-3B	900.5	54.16	93.99%
		Average % Time Savings per Request:	91.85%

. Only the completed requests (picked up and delivered) in both approaches are considered in the computation of turnaround time for direct comparison. Hence, 11224-1B to 3B is reflected. In the old framework, the turnaround time per reference number is measured by the total elapsed time from the origin to its delivery point (in Table 6) plus 12 h, since the earliest delivery in the existing method is the next day. On the other hand, same-day delivery can be achieved with the new framework since the requests are received through the app and automatically dispatched to the available driver within the delivery area, unlike in the old method where all picked-up goods needed to be dropped off first at the logistics head office for dispatch scheduling by the person in charge. With this development, the turnaround time is measured by the total elapsed time between the pick-up and delivery points on the trip, as shown in Table 7.

In terms of the order of pick-up points, in Table 6 and Table 7, they only differ between the third and fourth points, which reflects that the familiarity of the assigned driver or the handling dispatcher with the test locations is coincidentally high since these are within the metro area only. However, this is not always the case. Other times, the integrated waypoint optimization algorithm offered by this study gains significance in the sense of managing the route for the driver automatically, thus reducing the elapsed time between test points, since the following have been potentially eliminated from the conventional approach: (1) coordination between the driver and the dispatcher/customer in terms of customer and package detail verification, (2) searching for directions to the next point either manually or with the aid of existing navigation apps, and (3) coordination with the contact person of the following pick-up point.

3.2. Distance Traveled, Fuel Expenses, and CO₂ Emissions

Table 10. Summary of distance traveled from previous location using old and new frameworks.

			Old Framework		New Framework
Job Request No.	Service Location	Service (P/D)	Reference No.	Distance Traveled from Previous Location (km)	Reference No.	Distance Traveled from Previous Location (km)
Pick-Up Location	Mega Pacific Freight Logistics, Inc.	-		0		0
0	Br. Andrew Gonzales Hall	P	11324-1B	15.5	11324/11224-1B	15.5
1	University Pad Residences Taft	P	11324-2B	0.7	11324/11224-2B	0.7
2	Vista GL Taft by Vista Residences	P	11324-3B	2.5	11324/11224-3B	2.6
3	Robinsons Place Manila	P	11324-4B	0.4	11324-4B	1.5
4	Adamson University Main Building	D	11224-1B	2.1	11224-1B	1.6
5	SM Mall of Asia	D	11224-3B	7.9	11224-3B	6.9
6	SM City Bicutan	D	11224-2B	11.8	11224-2B	11.9
Drop-off Location	Mega Pacific Freight Logistics, Inc.	-	-	4.1		3

recorded the distance traveled from one point to another using the old and new frameworks. The rows are aligned with Table 6 and Table 7 in the previous section. Based on this, the total distance traversed per completed service request (picked up and delivered) was calculated and is summarized in

Table 11. Summary of distances traversed, fuel expenses, and carbon emissions for completed service requests using old and new frameworks.

	Distance Traversed (km)		Fuel Expenses (PHP)		CO2 Emissions (kg/veh-km)
Reference No.	Old	New	Old	New	Old	New	% Savings
11224-1B	17.60	1.60	143.00	13.00	4.96502	0.45137	90.91%
11224-2B	12.50	11.90	101.56	96.69	3.52629	3.35703	4.80%
11224-3B	10.40	6.90	84.50	56.06	2.93387	1.94651	33.65%
						Average % Savings per Request:	43.12%

. For the traditional approach, the distance travelled from the previous location, both for pick-up and delivery job requests, was considered in computing the total distance traversed per reference number since the two were performed on different days. On the other hand, only the distance traveled from the previous location for delivery job was considered as the distance traversed per reference number using the new framework since its corresponding pick-up job was performed simultaneously with a pick-up job for another customer on the same day. Hence, the driver saved time and fuel consumption by completing a single job for two customers with the exact same pick-up location but different delivery locations. Accordingly, the fuel consumption and carbon emissions were computed based on the distance traversed in Table 11. The carbon footprint was measured using the distance-based method considering a 0.454 CO₂ emission factor for light-duty trucks, as listed in the emission factors for the greenhouse gas inventory as of April 2022 [61,62]. The average percentage savings/decrease per service request computed for distance, fuel, and carbon using the new framework equated to 43.12%.

3.3. Projected Savings from Operating Expenses

The corresponding expenses using the old and new frameworks are summarized in

Table 12. Summary of operating expenses using old and new frameworks.

	Old	New
Particulars	Amount (PHP)	Amount (PHP)	Remarks
Labor Cost	1910.00	625.00	see Table 5 for labor rates using new framework
Driver’s Wage	1100.00	400.00	PHP 550 per day for 2 days
Helper’s Wage	810.00	225.00	PHP 405 per day for 2 days
Fuel Cost	650.81	316.01
Day 1 Job Requests	325.41	-	Pick-up of RF# 11224-1B to 3B
Day 2 Job Requests	325.41	-	Delivery of Day 1 RF# and Pick-up of RF#11324-1B TO 4b
Maintenance Cost	-	150.00
Total Expenses:	2560.81	1091.01	57.40% savings

. The expenses considered were the labor cost from the driver’s and helper’s wage rates and the fuel cost incurred from the total consumption throughout the trip. Two days’ worth of fees were computed for the labor cost in the old method since delivery per reference number cannot be carried out on the same day as the pick-up. Hence, the fuel consumption doubled due to the resulting two-way trip performed per service request. For the computation of labor, fuel, and maintenance costs in the new framework, the freight matrix in Table 5 was used as the reference. A total of 57.40% was reduced in the operating expenses using the new framework.

3.4. Projected Revenue and Profit from Completed Service Requests

Revenues that can be generated from service requests are based on the corresponding freight charges. As discussed in Section 2.7 and Section 2.8, charges apply dependent on the two-way distance in km of pick-up and delivery points per reference number (see Table 3 for the test scenario). It can be seen from

Table 13. Summary of revenue and profit using old and new frameworks.

Reference No.	Status	Distance in km (2-Way)	Old Freight Charge (PHP)	New Freight Charge (PHP)
11224-1B	Complete	11.4	1600	1000
11224-2B	Complete	49.4	2000	3300
11224-3B	Complete	20.8	1700	1600
11324-1B	For Next-Day Delivery	-
11324-2B	For Next-Day Delivery	-
11324-3B	For Next-Day Delivery	-
11324-4B	Complete	51.8	2100	3500	% Increase
		Total Revenue:	7400	9400	27.03%
		Total Expenses:	2560.81	1091.01
		Total Profit:	4839.19	8308.99	71.70%

that four out of seven requests were completed, since the three requests with delivery points outside the metro were considered for next-day delivery from the pick-up date. Hence, freight charges are not included in the computation of business earnings yet. The corresponding charges are indicated following the freight rates for the old and new frameworks (Table 4 and Table 5). It can be seen that freight charges using the new framework are lower than the old ones at shorter distances and higher at longer distances. This means that services over shorter distances will be more attractive to customers. However, the convenience offered by using the app means the higher freight charges at longer distances are still on par. In addition, longer distances have become more serviceable since the rates for drivers and helpers are made proportional to the distance. The use of the new framework yielded a 27% increase in revenue and, taking the total expenses from Table 12 into account, resulted in a profit increase of 72%.

3.5. User Acceptance

A survey was conducted to assess the users’ satisfaction with the developed app. A questionnaire was disseminated for users to rate the app on a five-point scale based on various categories, as listed in

Table 14. Summary of results gathered from user-acceptance survey.

Category	Survey Item	Mean Score
User Experience	The application software is user-friendly.	4.1
	The interface of the application software is intuitive and easy to navigate.	4.0
	Features on the application software will significantly improve workflow efficiency.	4.4
Maintenance of Vehicles	The application software will help lessen the strain on vehicles.	4.6
	The application software will contribute to extending the lifespan of the vehicles in the fleet.	4.7
Safety	The application software will positively influence the safety of the driver and helper during operations.	4.4
	The route optimization will reduce fatigue and enhance the safety of the work environment since it will lessen the worry of manually planning the route ahead.	4.8
Seamless Navigation	I am satisfied with the accuracy and reliability of the navigation system within the application software.	4.4
	The navigation feature will help avoid delays and optimize delivery routes effectively.	4.5
Optimization in Operations	The application software has significantly reduced the manual effort required by dispatchers and drivers in logistics operations.	4.8
	Automated processes like order processing and inventory management will improve overall efficiency.	4.9
Seamless Integration	The application software seamlessly integrates with existing systems (return trips and consolidated shipping) in operations.	4.1

, where 5 means very satisfied, 3 means neutral, and 1 means very dissatisfied. The mean score from all of the answers received was computed per item and is indicated in the mean score column of Table 14. It can be observed that the survey results were mostly positive, with all mean scores ranging between 4 and 5, regarding the application software. The application will greatly benefit the company, especially the logistics operations, and logistics staff can easily use the application software after being taught how to use it. The most notable part of the survey results was the response of the drivers, who gave a positive response, especially on the driver app, as they would no longer need to manually plan their routes and would avoid traffic as much as possible. The results indicate that the software application based on the proposed GIS-based framework can positively impact the user experience, maintenance of vehicles, and safety, as well as allow for seamless navigation, operation optimization, and seamless integration.

4. Conclusions and Recommendations

In this study, we designed, developed, implemented, and built upon a small-scale application software for a dispatcher management system following the proposed GIS-based framework. The application programming interface of Google Maps was integrated to perform waypoint order optimization with the aid of parameters from Directions, Distance Matrix, and polyline utility. Next, return trip node-based order gathering within a small-scale-area application was established. After that, application software for multiple roles/users such as an admin, dispatcher, driver, and customer was also developed. Lastly, the operations with the application software were compared to operations without the application software, and the results showed significant improvements and cost savings for clients based on different success metrics.

The test results using the new framework show it has promising value as it can reduce the travel time of the driver by optimizing the order of pick-up addresses while also considering the live traffic information in Google. The application also provides a significant improvement for the company as it can allow return trips without spending much time planning them (with the click of a button). The company can also benefit from the time savings that the app provides due to its efficient routing. The role of technology in easing the work burden of dispatchers is evident in this application. Without the quick and efficient planning of the application software, dispatchers would have to spend an enormous amount of time planning the drivers’ trips to optimize their paths, and they would do so without live traffic information [63]. The improved delivery efficiency and reduction in operational costs brought by the app can offer significant value to the company because of the shorter time that the vehicle will spend on the road and the potential profits for the company from using the application. The positive remarks of the respondents show the potential that this application can bring to a business, covering its essential operations, starting from the dispatcher and driver and extending up to the customer. The comments and suggestions of the respondents can be taken into consideration during the second phase of this study or by other researchers who opt to enhance our work in the future.

Implementing the system on other platforms such as iOS and web, in addition to the current mobile application, is suggested to serve a more extensive variety of users and devices. For seamless integration with the existing Enterprise Resource Planning (ERP) systems, it is recommended that the dispatch status should be repackaged into a callable function or API so that SMEs and even larger e-commerce businesses may consider the software as their 3PL software solution. The dispatch route function can also be converted into an API so that existing 3PL companies can consider integrating this with their systems, to aid their drivers in finding the optimal waypoint for their multiple deliveries throughout the day. Moreover, we recommend testing the system in various settings like different types of vehicles, including electric vehicles, for value-added sustainability, increasing the simultaneous trip requests, and considering test locations in less familiar areas or outside the metro, as well as longer distances between pick-up/delivery points. It is also suggested that the research field would benefit from revisiting the ways cargo spaces can be optimally utilized to reduce courier activities on the road. Finally, implementing a feedback feature in the program might also be beneficial since it will support a continuous cycle of development and refinement of the application by allowing the users to provide input and suggestions for system improvement.

Overall, we were able to achieve our objectives and contribute positively to making operations for small and medium-sized third-party logistics businesses more sustainable in the future.