Delivering Goods Sustainably: A Fuzzy Nonlinear Multi-Objective Programming Approach for E-Commerce Logistics in Taiwan

Abstract

With the booming development of e-commerce, the importance of controlling carbon emissions has become increasingly prominent in Taiwan. This study explores the trade-offs among time, cost, quality, and carbon emissions (TCQCE) in e-commerce logistics. Will carbon emissions mitigation lead to decreased logistics efficiency and increased costs? This article differs from other studies that use precise numbers and linear model situations. This study adopts fuzzy theory, nonlinear methods, and multi-objective programming models closer to the actual situation to study the decision-making between delayed logistics delivery times and reduced carbon emissions. This article also uses Project D as a case to enhance readers’ understanding of decision-making methods in real-life e-commerce logistics cases. The results show that extended delivery times could significantly reduce carbon emissions, ranging from 5259.31 to 419,199.60 tons, and reduce delivery quality under the 90.00% threshold and even under 75.25%. Extending delivery times is a viable business strategy, particularly by extending delivery to push carbon reduction policies to minimize environmental impact. However, consumer acceptance is crucial, as consumers willing to embrace longer wait times can significantly contribute to emission mitigation and support businesses committed to sustainability. This research uses a fuzzy nonlinear multi-objective programming model (FNMOPM) to contribute novel time management to mitigate carbon emissions. Moreover, this study uses a fuzzy and nonlinear approach to fill in the gaps of previous research to balance the efficiency and carbon emission mitigation goals of ESG (environmental, social, and governance) principles. The framework presented in this article solves the complex trade-off situations in the TCQCE issues. This article provides practical, actionable guidance for decision-making regarding sustainable e-commerce logistics, instilling confidence in its implementation.

1. Introduction

As the e-commerce sector expands at an unprecedented pace, the need to rein in carbon emissions has become increasingly acute. Mohammad et al. [1] have underscored the pressing need for additional delivery vehicles as e-commerce goods surge. Turkensteen and Heuvel [2] have urged e-commerce corporations to invest in logistics fleets to enhance service quality and reduce delivery times. Abdelhamid’s [3] research found that the speed of order fulfillment and shipping are among the most critical factors affecting customers’ repurchases on online platforms. However, while streamlining logistics, Ullah [4] observed that these efforts inadvertently lead to increased carbon emissions, directly contradicting the environmental protection principles of ESG (environmental, social, and governance) principles. Taiwanese e-commerce companies have raised the issue of whether delayed delivery can effectively reduce carbon emissions and limit fleet expansion. Therefore, it is essential to design comprehensive solutions to trade-off problems that involve time, cost, quality, and carbon emissions (TCQCE).

Wygonik and Goodchild [5] have played a pivotal role in understanding trade-off problems, particularly those of cost, quality (vehicle routing), and carbon emissions. Their research model enhances quality while reducing cost and carbon emissions, underscoring the importance of these factors in addressing the complex trade-off issue. Romeiro, Fransen, and Lucon [6] highlighted that the primary means of reducing greenhouse gas emissions are primarily in the transport, industry, and power generation sectors. Raj et al. [7] identified lack of visibility (LV), low-efficiency levels (LELs), and unpredictability (UE) as the top three risk barriers in logistics. Specifically regarding LV, Garvey, Iyer, and Nash [8] found that carbon emission mitigation could reduce a company’s exposure to future greenhouse gas (GHG) regulations or taxes. However, research suggests that carbon taxes have a minimal impact on industries like the internet and business services. The most significant effect remains on the size of the delivery fleet. For LEL, Lähdeaho and Hilmola [9] pointed out that in the logistics industry, vehicle route optimization is crucial to running competitive operations, as optimized routes and times can reduce operating costs and improve service quality. However, Alves de Araújo et al. [10] found that for UE, a multi-criteria decision support system utilizing a fuzzy analytic hierarchy process (FAHP) to address service expectations for last-mile delivery is consistent with the widespread use of transport generating carbon emissions. Still, increases in e-commerce sales may increase greenhouse gas emissions and harm the environment in urban areas.

The booming e-commerce sector has fueled logistics industry growth, but this comes with a significant environmental cost and rising carbon emissions. Traditional logistics models, prioritizing efficiency over sustainability, often overemphasize reduction costs and shortened shipping times, as shown in previous studies. This approach has led to larger delivery fleets and a sharp increase in the carbon footprint. This research proposes a multi-objective programming model that tackles the trade-off between efficiency (cost and time) and environmental impact (carbon emissions). The model delves deeper into the balance problem of carbon reduction and efficiency in e-commerce logistics using nonlinearity. This novel solution aims to reduce carbon emissions while maintaining logistics quality, achieved through nonlinear multi-objective programming. By embracing this approach, logistics companies can become leaders in sustainable development, actively integrating environmental responsibility into operations.

Many scholars have proposed using diverse logistics transportation methods to curtail carbon emissions. Heshmati et al. [11] proposed new green technologies, such as electric bicycles and cars, to replace existing delivery tools and minimize e-commerce’s environmental footprint. Mangiaracina et al. [12] investigated the use of lockers and crowdsourcing logistics to diminish carbon emissions. Gong [13] proposed that governments can promote sustainable practices through policies incentivizing businesses to extend delivery times, such as tax breaks or subsidies. Iwan et al. [14] focused on adopting electric vans for last-mile delivery to solve air pollution and noise impacts.

The studies mentioned above signify an escalating academic trend toward discovering innovative and sustainable solutions for e-commerce logistics. Introducing green technologies and novel logistics strategies aligns with the carbon emission mitigation goal of ESG principles and contributes to broader environmental conservation goals. However, the crucial challenge lies in implementing these sustainable practices without escalating costs. This study commenced with time management using nonlinear multi-objective programming to address the waiting days of shipping times to reduce carbon emissions, thereby maintaining the convenience of e-commerce and accomplishing the carbon emission mitigation goal of ESG principles.

What is the optimization methodology regarding TCQCE in e-commerce? In 2024, Kim et al. [15], aiming at the problem of closed-loop supply chain network design (CLSCND), proposed a multi-objective mixed-integer linear programming (MMILP) model. Its objectives include minimizing total costs, minimizing carbon dioxide emissions, and maximizing employment opportunities. The results showed that using the MMILP model can provide meaningful parameters for strategic decisions of business companies and managers.

Rezakhanlou and Mirzapour Al-e-Hashem [16] leveraged a multi-objective linear programming model to analyze the optimal configuration, pricing, and transportation decisions within an e-commerce supply chain. Rezakhanlou et al.’s model encompasses traditional stores, e-commerce home delivery, and dual-channel network structures under configuration uncertainty. The approach integrates environmental sustainability by minimizing carbon dioxide emissions associated with transportation and production processes, ensuring adherence to environmental regulations. Each distribution pipeline offers multiple transportation modes with varying costs and carbon footprints. The authors demonstrate the proposed model’s effectiveness and the solution method’s practicality through numerical examples. These examples reveal that e-commerce companies implementing dual-channel supply chains can significantly improve profitability and environmental performance.

Kim et al. [15] and Rezakhanlou et al. [16] focused on a linear model. There is still a gap between linear models and the actual environment. Parilina, Yao, and Zaccour also evaluated the trade-off issue between sustainability and efficiency [17]. Nonlinear programming offers a more robust approach by representing the nonlinear interactions among variables. Furthermore, real-world uncertainty factors and environmental dynamics can be imprecise. The fuzzy theory addresses these uncertainty factors by incorporating degrees of membership instead of relying solely on strict binary values. Therefore, integrating fuzzy theory and nonlinear multi-objective programming offers a robust framework for addressing trade-offs in decision-making.

Driven by the challenge of complex environmental situations, this study proposes a fuzzy nonlinear multi-objective programming model to tackle the trade-off between the logistics of TCQCE and carbon emission mitigation of ESG principles in e-commerce operations. This fuzzy nonlinear multi-objective programming offers practical solutions for e-commerce companies. This approach bridges existing research gaps by incorporating fuzzy theory to account for ecological uncertainty and utilizing nonlinear models for a more comprehensive representation. Furthermore, this model provides valuable delivery guidance for policymakers, regulators, and stakeholders, ultimately promoting sustainable development within the e-commerce industry.

2. Literature Review

In 2019, Olsson, Hellström, and Pålsson [18] researched the final stage of supply chain operations, commonly known as last-mile logistics. The scholars identified the phase as the process’s most expensive, least effective, and environmentally harmful part. Olsson et al. analyzed 155 articles, categorizing them into five themes: (1) merging trends and technologies, (2) operational optimization, (3) supply chain structure, (4) performance measurement, and (5) policy. These findings served as the foundation for our current research.

In response to these discoveries, this paper introduces a model that examines the trade-offs among time, cost, quality, and carbon emissions for operational optimization. These trade-offs emphasize the significance of delivery time considerations, which influence quality, cost, and carbon emissions. The model aims to support the last-mile logistics aligned with the carbon reduction targets outlined in the carbon emission mitigation goals of the ESG principles, a vital aspect of the sustainable development goals (SDGs) framework.

Edwards et al. [19] investigated the economic consequences of failed deliveries and the associated carbon footprint. The research suggests that employing collection–delivery points (CDPs), such as supermarkets, post offices, and train stations, can significantly mitigate environmental burden. The U.K. Royal Mail’s CDP system is the most cost-effective solution, generating only 13% of the carbon dioxide emissions produced by traditional failed deliveries. The approach offers consumers a convenient and environmentally friendly alternative.

Jiang et al. [20] explained that due to the rapid growth of online shopping package volume, many secondary deliveries have high wasted costs. The study proposed a variant variable neighborhood descent (VVND) algorithm to solve the problem of a mixed customer pickup (CP) and home delivery (HD) delivery model. The VVND algorithm demonstrated superior performance and competitiveness compared to existing methods.

Chen, Qiu, and Hu [21] tackled the loaded freight location-routing problem with full truckloads (LRPFT) to minimize transportation routes for carbon emissions. The work focused on the mathematical model of attaining the multiple objectives of total cost and environmental impact minimization, combining the strengths of the non-dominated sorting genetic algorithm-II (NSGA-II) and the tabu search (TS) heuristic for the novel multi-objective hybrid method, NSGA-II-TS. Numerical simulations validated the effectiveness of the hybrid approach, while a real-world case study demonstrated the trade-off between the total cost and carbon dioxide emissions in LRPFT scenarios.

Li et al. [22] studied how waste recycling supply chain logistics can reduce carbon emissions, including facility capacity, transportation costs per unit of product per kilometer, landfill costs, unit carbon for penalty costs, and carbon incentive amounts. The goal of the proposed fuzzy optimization model is to minimize the total cost of the network and the sum of carbon rewards and penalties when selecting facility locations and transportation routes among network nodes. Changes in the penalty cost and incentive amount per unit of carbon emissions significantly impact the carbon penalty revenue and total expenses.

Due to mutual competition, e-commerce companies aim to deliver fast shipments. In addition to logistics routes and CDP, e-commerce companies constantly expand delivery fleets to shorten their delivery times or purchase electric vehicles to reduce pollution and achieve customer satisfaction. Therefore, based on the above research, scholars have proposed methods to reduce carbon emissions and costs under sustainable e-commerce logistics and reduce costs while reducing carbon emissions. The above three papers utilized operational optimization methods (such as CDPs and LRPFT), making optimization a pivotal approach to reducing carbon emissions in sustainable e-commerce logistics.

Zhao et al. [23] investigated the objective optimization problem concerning a hybrid fleet composed of conventional and electric vehicles within a specified time window. Electric cars, characterized by low greenhouse gas emissions and quiet operation, will increase the number of electric vehicle fleets that offer good service. Nevertheless, the substantial initial cost and restricted driving range of electric cars limit their potential to supplant traditional fleets entirely. The study employed the non-dominated sorting genetic algorithm II (NSGA-II) and a linear weighting method. The findings underscore the inherent trade-offs in transitioning from traditional to electric fleets.

Banerjee et al. [24] designed a same-day delivery system. The researchers utilized a Markov decision process to investigate a setting with multiple vehicles dispatched once daily, facilitating the analysis of the marginal benefit of increasing the fleet size. The results show that the Markov decision-making process can effectively realize the plan’s marginal benefit.

Both studies mentioned above emphasized the importance of trade-offs and optimization methods in decision-making to ensure efficiency, cost optimization, and utilization of electric fleet technology to reduce carbon emissions. However, e-commerce companies invest in many fleets to shorten delivery time. Companies should use time management methods to reduce their vast investments and complete the purpose of goods delivery while improving their carbon emissions. Therefore, using a time management approach, this study fills the research gap in low-cost investments and carbon emission reduction.

By analyzing online customer data, Amorim and DeHoratius [25] demonstrated that consumers prioritize delivery attributes beyond the order fulfillment speed. The COVID-19 pandemic changed the way shoppers engage online. As of May 2020, 41% of U.S. online grocery shoppers were first-time users of such services. Some customers prioritize being able to choose the delivery day over speed, and others might be willing to forgo speed in return for a precise delivery time window.

Cheah and Huang [26] elucidated that e-commerce shopping platforms typically offer multiple shipping options to accommodate consumer preferences. Faster delivery can be detrimental to the environment. Cheah and Huang carried out a transportation preference survey to assess the influence of carbon labels on consumers’ transportation choices. The findings indicated that 55% of the 188 respondents were ready to forgo delivery speed to select less carbon-intensive alternatives based on the cost, speed, and carbon emission values of the scenarios presented in the survey.

In 2024, Biancolin and Rotaris [27] highlighted the significant environmental impact arising from the rapid growth of e-commerce. The study investigated the issue using multinomial logit (MNL) and mixed multinomial logit (MMNL) models, analyzing data from a sample of 1204 Italian consumers. The findings revealed that consumers were willing to pay a surcharge of EUR 0.88 to offset the environmental impacts, potentially mitigating the ecological footprint of last-mile logistics. Consumers are willing to pay an additional EUR 0.17 to support reforestation initiatives to help mitigate the environmental impact of their purchase. Furthermore, the study identified a segment of environmentally conscious consumers willing to accept extended delivery times in exchange for discounts ranging from EUR 0.20 to 0.80.

Consumer behavior suggests a willingness to wait longer for orders, potentially leading to reduced carbon emissions. Cui et al. [28] surveyed the causal effects of retailers’ delivery speed commitments on customer behavior and business performance. Faster delivery can ensure more customer satisfaction. The findings showed that retailers can use data to customize the days of delivery. To reduce the carbon emissions generated by delivery, consumers can wait 3–5 days, making it easier to arrange transportation equipment, routes, CDPs, or time windows to achieve the carbon emission mitigation goals of ESG principles.

Regarding ESG principles, Shu and Tan [29] highlighted the critical role of ESG performance in driving sustainable corporate development. An examination of a dataset of China’s listed industrial enterprises from 2010 to 2019 reveals that carbon control policies pose significant challenges to corporate ESG performance. Negative impact arises from including carbon emissions as an evaluation metric for financing constraints and bank loan costs. Companies that prioritize mitigation policy risk and actively engage in ESG initiatives can achieve better outcomes.

Regarding the carbon emission mitigation goals of ESG principles, Baratta et al. [30] observed a gradual integration of sustainable development goals within corporate strategies alongside the incorporation of ESG considerations. Utilizing bibliometric analysis, they demonstrated a growing acknowledgment of the interconnectedness between future factories, industrial operations, and carbon emissions. Furthermore, they advocated for adopting an ESG-centered strategic framework concerning key performance indicators (KPIs).

Long and Feng [31] emphasized the critical importance of balancing economic growth with carbon emissions to address the challenge of irreversible climate change. Using a spanning data analysis from 41 countries from 1990 to 2020, they elucidated that the strength of environmental policies correlates with a more pronounced suppression of greenhouse gas emissions through enhanced ESG performance.

The above literature on ESG performance illustrates its critical position in corporate strategies and highlights the importance of policy and environmental factors in ESG principles. Therefore, these research results indicate that the carbon emission mitigation goals of ESG principles provide essential reference and guidance for implementing sustainable development goals.

According to all the above literature, consumers outside companies are willing to pay more to achieve carbon reduction. Regarding internal carbon reduction methods within enterprises, carbon emissions can be improved through delivery routes and installing receiving equipment, and consumers are tolerant of extending the number of days it takes for goods to arrive. However, how can e-commerce companies maintain a service quality above 90.00% without incurring additional vehicle costs? What is the number of “T” days needed (in TCQCE) to reduce carbon emissions? Given the uncertainty of the logistics environment, the difference in delivery time can be resolved by applying Bellman and Zade’s [32] fuzzy theory. Ivohin, Gavrylenko, and Ivohina [33] implemented a shift from a fuzzy to a parametric formulation in the form outlined by Bellman–Zade to create fuzzy numerical sets that formalize the “fast” and “slow” progression of time. This established method offers a potential solution to the issue of formalizing an individual’s subjective perception of time in processes that involve human participation. In addition, the multi-objective method tackles the challenge of expressing fuzzy time and the trade-off problem discussed by Chen, Qiu, and Hu [21] in the previous literature. For the application of multi-objective methods, Wang et al. [34] used the nonlinear mathematical programming model of Zimmermann [35] to solve the time, cost, and quality (TCQ) trade-off problem in solar photovoltaic station construction in logistics warehouses to achieve the sustainability goals of ESG principles.

According to the above literature, this study introduces fuzzy and nonlinear parameters to represent uncertainties in supply chain factors, making the decision-making process more robust to real-world variations. Establishing effective time management is crucial for achieving the goal of carbon reduction. This research integrates the fuzzy theory, nonlinear methods, and a multi-objective programming model (FNMOPM) to address the trade-off problem of total cost, quality, and carbon emissions (TCQCE). Notably, previous research should utilize the fuzzy theory and nonlinear methods. The lack of these methods limits the ability to capture the nuances of real-world situations. This study bridges the gap by employing fuzzy and nonlinear methods to enhance the rigor and applicability of the research results.

3. Methodology

3.1. Research Structure

This research aims to propose a FNMOPM that incorporates time management to address the trade-offs between TCQCE and delivery time in sustainable e-commerce logistics. From a consumer’s perspective, the generally acceptable delivery time ranges from 1 to 3 days. To reduce costs and carbon emissions, e-commerce companies can incorporate an additional two-day delivery window into the policy of logistics arrangements, treating each delivery task as a project. Therefore, proactive planning and evaluation are essential in complex environments within the research framework. The critical question is how to extend the delivery time in a fuzzy environment to resolve the TCQCE trade-off problem in logistics delivery and ultimately achieve the goal of reducing carbon emissions, as illustrated in Figure 1.

3.2. Case Assumptions

In logistics transportation, time is critical, and costs escalate when fast delivery and high quality are necessary. Due to time constraints, carbon emissions also rise. This study puts forth a series of hypotheses to illustrate these scenarios:

Hypothesis  1:

Implementing longer shipping times can result in reduced costs.

Hypothesis  2:

Delayed shipping times result in lower carbon emissions.

Hypothesis  3:

Quality requirements impact cost increases in trade-offs.

These assumptions abide by the critical concept that time significantly reduces carbon emissions. Effective time management can apply the carbon emission mitigation goals of ESG principles to enhance carbon management.

3.3. Case Methodology

Drawing upon the research methodologies of many academics, this investigation culminated in creating a comprehensive computational model for FNMOPM. This paper’s approach amalgamates Chen, Qiu, and Hu’s [21] multi-objective model connection with Bellman and Zade’s [32] fuzzy theory along with its parametric transition by Ivohin, Gavrylenko, and Ivohina [33], while simultaneously utilizing Zimmermann’s [35] nonlinear mathematical programming. Furthermore, it incorporates Dong and Shah’s [36] method of fuzzy computation and the α-cut defuzzification mode proposed by Leandry, Sosoma, and Koloseni [37].

Each project is an independent task. Adjust each project’s duration with slack (buffer time) in each schedule to accommodate potential delays without impacting delivery times. The following describes the development of subsequent models:

$$Max{ S}_{finish}$$

(1)

This maximizes the scheduling time of the delayed delivery status in the independent project.

$${Min C}_i$$

(2)

This minimizes the scheduling cost of the delayed delivery status in the independent project.

$$Max Q_i$$

(3)

This maximizes the scheduling quality of the delayed delivery status in the independent project.

The restrictions are:

$$s_{Fi}+{Lag}_{Fi}\le s_{Fi}+t_{Fi}$$

(4)

The total time of independent projects cannot be later than the independent projects’ planned start time plus delay time.

$$s_{Fi}\ge 0$$

(5)

This is the non-negative bounds on the start time of each fuzzy project.

$$t_{Fi}\in \mathbb{R}$$

(6)

The start time of each fuzzy project is a positive integer.

$$t_{normal,Fi}\le t_{Fi}\le t_{delay,Fi}$$

(7)

$$c_{delay,Fi}\le c_{Fi}\le c_{normal,Fi}$$

(8)

$$q_{delay,Fi}\le q_{Fi}\le 100\%$$

(9)

The calculation includes the fuzzy time cost ($C_{T,Fi}$), the direct fuzzy cost ($C_{Fi}$), the fuzzy quality cost ($C_{Q,Fi}$), and the fuzzy direct carbon emissions (${CE}_{T,Fi}$). This method is consistent with the groundbreaking research identifying the inherent trade-off between time and quality. As a result, the original linear relationship between time and cost becomes a nonlinear function. Carbon emissions are the subtractive component of the trade-off equation. Based on groundwork, Zimmermann proposed a mathematical model that sets $a_{Fi}$ as a constant, allowing for the following calculation.

$${CE}_{T,Fi}=a_{Fi}\times (t_{Fi}{)}^2\times {CE}_{Fi}+b_{Fi}$$

(10)

$$a_{Fi}=\frac{c_{normal,Fi}-c_{delay,Fi}}{t_{normal,Fi}^2-t_{delay,Fi}^2}$$

(11)

$$b_{Fi}=\frac{{c_{delay,Fi}\times t}_{normal,Fi}^2-c_{normal,Fi}\times t_{delay,Fi}^2}{t_{normal,Fi}^2-t_{delay,Fi}^2}$$

(12)

$$\begin{gathered} C_{Q,Fi}=\left[(\frac{∆C_{Q_{delay,Fi}}-∆C_{Q_{normal,Fi}}}{t_{delay,Fi}-t_{normal,Fi}})\times \left(t_{Fi}-t_{normal,Fi}\right)+∆C_{Q_{delay,Fi}}\right] \\ \times \left(q_{Fi}-Q_{T,Fi}\right) \end{gathered}$$

(13)

$$Q_{T,Fi}=q_{delay,Fi}+\frac{q_{delay,Fi}-q_{normal,Fi}}{t_{delay,Fi}-t_{normal,Fi}}\times \left(t_{Fi}-t_{delay,Fi}\right)$$

(14)

The definitions of the above symbols are as follows:

$Fi:$ This is the number of each fuzzy project.

$t_{Fi}:$ This is the duration of each fuzzy project.

$q_{Fi}$: This is the quality ratio of each fuzzy project.

$c_{Fi}$: This is the cost amount of each fuzzy project.

$t_{delay,Fi}$: This is the duration time of each fuzzy project in the delayed status.

$t_{normal,Fi}$: This is the duration of each fuzzy project in the normal status.

$c_{delay,Fi}$: This is the direct cost of the delayed working status of each fuzzy project.

$c_{normal, Fi}$: This is the direct cost of the normal work status of each fuzzy project.

$q_{delay,Fi}$: This is the quality ratio of each fuzzy project in the delayed status.

$q_{normal,Fi}$: This is the quality ratio of the normal status of each fuzzy project.

${CE}_{T,Fi}$: This is the carbon emissions status of each fuzzy project.

$C_{Q,Fi}$: This is the calculation of the cost of quality ratio for each fuzzy project.

$Q_{T,Fi}$: This is the calculation of the quality of time ratio for each fuzzy project.

$∆C_{Q_{delay,Fi}}$: This is the unit cost of the quality ratio variation of each fuzzy project in the delayed status.

$∆C_{Q_{normal,Fi}}$: This is the unit cost of the quality ratio variation of each fuzzy project in the normal status.

In fuzzy theory, the α-cut method is a leading technique for ranking fuzzy numbers. Fuzzy numbers represent sets of real numbers, each element holding an associated membership degree that reflects its level of belonging to the set. The membership function formalizes the concept by assigning a numerical degree of membership to each element within the fuzzy set, which symbolizes a fuzzy number characterized by two membership functions, $f^{a_{FiL}}$ and $f^{a_{FiR}}$, corresponding to its left and right membership functions. Assuming the existence of the equations’ inverse functions, $g^{a_{FiL}}$ and $g^{a_{FiR}}$, for $f^{a_{FiL}}$ and $f^{a_{FiR}}$, the article defines the left integral value, $I^L\left(a_{Fi}\right)$, and the right integral value, $I^L\left(a_{Fi}\right)$, of $a_i$ as outlined in Equations (16) and (17), presented by Leandry, Sosoma, and Koloseni [37] and Liou and Wang [38], respectively. Furthermore, the mathematical operations on two positive fuzzy numbers denoted as A and B can be expressed through fuzzy arithmetic as follows: fuzzy addition: (A$\oplus$B)^α = [A_l^α + B_l^α, A_u^α + B_u^α]; fuzzy subtraction: (A$\ominus$B)^α = [A_l^α − B_u^α, A_u^α − B_l^α]; fuzzy multiplication: (A$\otimes$B)^α = [A_l^αB_l^α, A_u^αB_u^α]; and fuzzy division: (A$\oslash$B)^α = [A_l^α/B_u^α, A_u^α/B_l^α]:

$$(a_{Fi}{)}_{\alpha }=\left[(a_i{)}_{\alpha }^L, (a_i{)}_{\alpha }^U\right]$$

(15)

$$I^L\left(a_{Fi}\right)=\frac{1}{2}\underset{k\to \mathrm{\infty }}{\lim }\left\{\sum _{j=1}^k\left[{g_{{\alpha }_{Fi}}}^L\left({\alpha }_{Fj}\right)+{g_{{\alpha }_{Fi}}}^L\left({\alpha }_{Fj-1}\right)\right]∆{\alpha }_{Fj}\right\}$$

(16)

$$I^R\left(a_{Fi}\right)=\frac{1}{2}\underset{k\to \mathrm{\infty }}{\lim }\left\{\sum _{j=1}^k\left[{g_{{\alpha }_{Fi}}}^R\left({\alpha }_{Fj}\right)+{g_{{\alpha }_{Fi}}}^R\left({\alpha }_{Fj-1}\right)\right]∆{\alpha }_{Fj}\right\}$$

(17)

This paper introduces a novel logistics delivery project-scheduling paradigm explicitly targeting deadline optimization. By incorporating a granular analysis of execution phases, the model maintains a dynamic TCQCE equilibrium for each project element. It facilitates the generation of an optimal, streamlined schedule while approximating the project cost. The model prioritizes maximizing completion time, curtailing expenses, bolstering quality metrics, and minimizing carbon emissions. The methodology leverages fuzzy programming implemented in Excel 2016 and Python 3.11 to simulate logistics delivery projects. First, enter the Excel data into the Python fuzzy program for processing. After generating the results, systematic tabulation allows for comprehensive project analysis.

4. Sample Problem and Results

This paper introduces a time management approach for decision-oriented delivery project administration. The approach employs fuzzy theory, α-cut defuzzification, nonlinearity, mathematical programming, and multi-objectivity to address the correlations between TCQCE and the carbon emission mitigation goals of ESG principles. This research provides a comprehensive technique that enables the decision-maker to convert fuzzy quantities into exact values.

4.1. Case Introduction

This study focuses on SASSWOOD Corporation’s efforts to reduce its carbon emissions and improve performance. Switzerland’s subsidiary SASSWOOD in Taiwan proposed a new logistics delivery policy, extending the arrival timeframe for goods additional 2 days. The case data are shown in Table 1.

To establish a FNMOPM, follow these solution steps:

Step 1: Refer to Figure 2’s network map.

Step 2: Assess the lengthening duration of each project. Lengthened time = delayed time ($t_{delay,Fi})$—normal time ($t_{normal,Fi}$). Determine each project’s delayed cost ($c_{delay,Fi}$) and normal cost ($c_{normal,Fi}$), and calculate the slope (${∆C}_{Fi}$).

Step 3: Establishing a fuzzy situation.

Step 4: Next, to account for varying degrees of uncertainty, researchers applied α-cuts to fuzzy numbers $(\widetilde{c_{LP}}{)}_{\alpha },$ resulting in definite numbers.

In each scenario, make the following assumptions:

Shipping delays: Goods may experience delays of up to 2 days.

Fuel consumption: According to the Transportation Research Institute of the Ministry of Transport of Taiwan, a 1.5-ton truck consumes fuel at 10 km per liter. The carbon emission coefficient for gasoline is 2.39 kg per liter. Thus, to determine the carbon emissions for a delivery process involving 8 h of daily operation and a distance of 100 km, employ the following formula:

Carbon emissions = Fuel consumption $\times$ Carbon emission coefficient $\times$ Driving distance

Carbon emissions = 10 $\times$ 2.39 $\times$ 100 $=$ 2390 kg $=$ 2.39 tons.

Carbon emission reduction: Each day of delay reduces carbon emissions by 10%.

Decision criterion: Prioritize the longest delivery time. If multiple options have the same delay, choose the delayed schedule over standard delivery options.

Quality threshold: Quality must remain above 90.00%.

Table 1. Case data.

Project	Time (Days)		Cost (TWD)		Quality (%)		Quality Cost (TWD/%)
	Normal	Delay	Normal	Delay	Normal	Delay	Normal	Delay
A	3	5	10,500	9500	94.00	90.00	1600	3200
B	2	4	7000	7600	96.00	92.00	2400	2800
C	2	4	7000	7600	96.00	92.00	2400	2800
D	4	5	14,000	9500	85.00	80.00	600	3200
E	5	0	17,500	17,500	80.00	80.00	3600	3600

Table 1. Case data.

Project	Time (Days)		Cost (TWD)		Quality (%)		Quality Cost (TWD/%)
	Normal	Delay	Normal	Delay	Normal	Delay	Normal	Delay
A	3	5	10,500	9500	94.00	90.00	1600	3200
B	2	4	7000	7600	96.00	92.00	2400	2800
C	2	4	7000	7600	96.00	92.00	2400	2800
D	4	5	14,000	9500	85.00	80.00	600	3200
E	5	0	17,500	17,500	80.00	80.00	3600	3600

According to Table 1’s data, each project is an independent path to meeting the logistics last-mile characteristics of the final stage from Olsson, Hell-ström, and Pålsson [18]. Let e-commerce and the delivery of goods to customers use the independent routes, as shown in Figure 2.

Figure 2. The network map includes independent projects A–E.

Figure 2. The network map includes independent projects A–E.

Next, the authors calculated each project’s time delay, unit cost slope, and unit quality slope, as illustrated in

Table 2. Each project’s delayed time, unit time cost, and unit time quality.

Project	Time (Days)		Cost (TWD)		Quality (%)		Delay Time (Days) (7) = (2) − (1)	Unit Time Cost (TWD/Day) [((4) − (3))/(7)]	Unit Time Quality (%/Day) [((5) − (6))/(7)]
	Normal (1)	Delay (2)	Normal (3)	Delay (4)	Normal (5)	Delay (6)
A	3	5	10,500	9500	94.00	90.00	−2	−1000	−2
B	2	4	7000	7600	96.00	92.00	−2	600	−2
C	2	4	7000	7600	96.00	92.00	−2	600	−2
* D	4	5	14,000	9500	85.00	80.00	−1	−4500	−5
E	5	0	17,500	17,500	80.00	80.00	-	-	-

Note: The * represented Project D extending one day. The cost decreased to a maximum of TWD 4500, and quality decreased by 5%.

. For Project D, the time delay was one day, reducing TWD 4500 in unit time cost. Simultaneously, the unit time quality also decreased by 5%. However, due to stringent quality requirements, the quality must exceed 90.00%. Consequently, quality costs will increase.

4.2. FNMOPM for TCQCE

Zimmermann [35] identified the time cost slope of the membership function as ${\mu }_{{\tilde{a}}_l}\left(c_i\right)$, defining the fuzzy set of $\widetilde{a_i}$ as $\left(\stackrel{~}{{{\mathrm{c}}_l}_{\alpha }}\right)=\left\{c_i|{\mu }_{\widetilde{c_l}}\left(c_i\right)\ge \alpha \right\}$. This interval represents all $(\widetilde{c_l}{)}_{\alpha }=\left[(c_i{)}_{\alpha }^L, (c_i{)}_{\alpha }^U\right]$ values associated with possibility $a_i$. Within this model, the fuzzy unit time cost slope $\widetilde{a_i}$ for each operation Project $(\widetilde{c_i}{)}_{\alpha }=\left[(c_i{)}_{\alpha }^L,(c_i{)}_{\alpha }^M, (c_i{)}_{\alpha }^U\right]$ is modeled as a triangular fuzzy number. As an illustrative example in Project A’s time fuzzy number, $(\widetilde{c_A}{)}_{\alpha }=\left[(c_A{)}_{\alpha }^L,(c_A{)}_{\alpha }^M, {(c}_A{)}_{\alpha }^U\right]$. Set the Project’s average numbers and then calculate the cube for $(c_A{)}_{\alpha }^L,(c_A{)}_{\alpha }^M, {(c}_A{)}_{\alpha }^U$. The next step is to set up to add 10% for an upper boundary and decrease 10% for a lower boundary, as shown in

Table 3. The MP fuzzy nonlinear variations of Projects A, B, C, D, and E.

Project	Time (Days)				Cost (TWD)				Quality (%)
	$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{L}}$	$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{M}}$		$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{U}}$	$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{L}}$		$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{M}}$	$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{U}}$	$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{L}}$		$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{M}}$		$({\mathit{c}}_{\mathit{i}}{)}_{\mathit{\alpha }}^{\mathit{U}}$
A	* 3.52	* 3.91		* 4.31	8992.49		9991.66	10,990.83	82.79		91.99		101.18
B	2.6	2.88		3.17	6566.30		7295.89	8025.48	84.59		93.99		103.38
C	2.6	2.88		3.17	6566.30		7295.89	8025.48	84.59		93.99		103.38
D	4.03	4.48		4.93	10,444.13		11,604.59	12,765.05	74.23		82.47		90.72
E	4.50	5.00		5.5	15,750.00		17,500.00	19,250.00	72.00		80.00		88.00
Project	Unit time cost							Unit time quality
	$(c_i{)}_{\alpha }^L$		$(c_i{)}_{\alpha }^M$			$(c_i{)}_{\alpha }^U$		$(c_i{)}_{\alpha }^L$		$(c_i{)}_{\alpha }^M$		$(c_i{)}_{\alpha }^U$
A	−896.994		−996.66			−1096.326		−1.791		−1.99		−2.189
B	538.191		597.99			657.789		−1.791		−1.99		−2.189
C	538.191		597.99			657.789		−1.791		−1.99		−2.189
D	−4036.455		−4484.95			−4933.445		−4.482		−4.98		−5.478
E	0		0			0		0		0		0

Note: The * represented Project A’s average is (3 + 5)/2 = 4, $(c_A{)}_{\alpha }^M=\sqrt[3]{3\times 4\times 5}=3.9149\approx 3.91,(c_A{)}_{\alpha }^L=3.5234\approx 3.52,(c_A{)}_{\alpha }^U=4.3063\approx 4.31$.

.

As shown in

Table 4. Defuzzification using α-cuts for time ranking values.

α	Project
	A	B	C	D	E
0	[3.52, 4.31]	[2.6, 3.17]	[2.6, 3.17]	[4.03, 4.93]	[4.5, 5.5]
0.2	[3.598, 4.23]	[2.656, 3.112]	[2.656, 3.112]	[4.12, 4.84]	[4.6, 5.4]
0.4	[3.676, 4.15]	[2.712, 3.054]	[2.712, 3.054]	[4.21, 4.75]	[4.7, 5.3]
0.6	[3.754, 4.07]	[2.768, 2.996]	[2.768, 2.996]	[4.3, 4.66]	[4.8, 5.2]
0.8	[3.832, 3.99]	[2.824, 2.938]	[2.824, 2.938]	[4.39, 4.57]	[4.9, 5.1]
1	[3.91, 3.91]	[2.88, 2.88]	[2.88, 2.88]	[4.48, 4.48]	[5, 5]
Time ranking	7.825	5.765	5.765	8.96	* 10

Note: The * represented Project E has the highest defuzzification value of 10.

, Projects A, B, C, D, and E have defuzzification values of 7.825, 5.765, 5.765, 8.96, and 10. However, it is essential to recognize that these values do not directly correlate with increased logistics delivery days. Instead, the numerical values serve as relative size indicators. This implies that the expected logistics delivery time for the other projects and the extended logistics delivery time result in the shipment’s arrival time falling exclusively within the extended delivery window. This can achieve the goal of carbon emission reduction. However, Project E’s logistics delivery time was initially set at five days. Consequently, the arrival time of the logistics deliveries can occur with minimal carbon emissions.

The defuzzification ranking (

Table 5. Each project performs defuzzification using α-cuts for cost ranking values.

α	Project
	A	B	C	D	E
0	[8992.5, 10,991]	[6566.3, 8025.5]	[6566.3, 8025.5]	[10,444, 12,765]	[15,750, 19,250]
0.2	[9192.3, 10,791]	[6712.2, 7879.6]	[6712.2, 7879.6]	[10,676, 12,533]	[16,100, 18,900]
0.4	[9392.2, 10,591]	[6858.1, 7733.6]	[6858.1, 7733.6]	[10,908, 12,301]	[16,450, 18,550]
0.6	[9592, 10,391]	[7004.1, 7587.7]	[7004.1, 7587.7]	[11,140, 12,069]	[16,800, 18,200]
0.8	[9791.8, 10,191]	[7150, 7441.8]	[7150, 7441.8]	[11,372, 11,837]	[17,150, 17,850]
1	[9991.7, 9991.7]	[7295.9, 7295.9]	[7295.9, 7295.9]	[11,605, 11,605]	[17,500, 17,500]
Cost ranking	19,983.32	* 14,591.78	* 14,591.78	23,209.18	35,000

Note: The * represented Projects B and C share the same logistics delivery time and minimal cost ranking values.

) reveals the cost ranking values for Projects A, B, C, D, and E as follows: A = 19,983.32, B = 14,591.78, C = 14,591.78, D = 23,209.18, and E = 35,000. Upon examining these ranking values, it becomes evident that the cost corresponds to the smallest ranking value. Although the costs for Projects B and C are the lowest, whether the carbon emissions are also the lowest remains to be determined through subsequent analysis.

As depicted in

Table 6. Defuzzification of each project using α-cuts for quality ranking values.

α	Project
	A	B	C	D	E
0	[82.79, 101.18]	[84.59, 103.38]	[84.59, 103.38]	[74.23, 90.72]	[72, 88]
0.2	[84.63, 99.342]	[86.47, 101.5]	[86.47, 101.5]	[75.878, 89.07]	[73.6, 86.4]
0.4	[86.47, 97.504]	[88.35, 99.624]	[88.35, 99.624]	[77.526, 87.42]	[75.2, 84.8]
0.6	[88.31, 95.666]	[90.23, 97.746]	[90.23, 97.746]	[79.174, 85.77]	[76.8, 83.2]
0.8	[90.15, 93.828]	[92.11, 95.868]	[92.11, 95.868]	[80.822, 84.12]	[78.4, 81.6]
1	[91.99, 91.99]	[93.99, 93.99]	[93.99, 93.99]	[82.47, 82.47]	[80, 80]
Quality ranking	183.975	* 187.975	* 187.975	164.945	160

Note: The * represented Projects B and C exhibit alignment and the highest quality ranking values.

, a high-quality ranking value signifies that the accuracy of the logistics delivery is relatively superior, resulting in shorter transit times for goods compared to the other projects. Consequently, the quality ranking value surpasses that of the other projects. Projects B and C consistently have commendable low-cost and high-quality performance. In organizational operations, low cost represents favorable performance characteristics. However, whether these projects also exhibit relatively low carbon emissions remains an important consideration. In addition to cost and quality, addressing low carbon emissions significantly affects the company’s overall sales image, as discussed by Deshmukh and Tare [39].

Analysis of the unit time cost revealed that the values of Projects A and D were negative. In this context, the unit time cost necessitated a minimum value. Project D will extend the logistics delivery time while effectively reducing costs. This underscores the inherent trade-off between time and cost within Project D. When time becomes a significant factor in cost control, which increases time efficiency, costs decrease. Therefore, time efficiency is an effective tool for managing costs. Project E manifests as a straight line within the triangular fuzzy number due to the absence of a significant variation in the number of days. The lower and upper bounds coincide, resulting in a data value of 0. As illustrated in

Table 7. Defuzzification of each project based on α-cuts for unit time cost ranking values.

α	Project
	A	B	C	D	E
0	[−897, −1096]	[538.19, 657.79]	[538.19, 657.79]	[−4036, −4933]	[0, 0]
0.2	[−916.9, −1076]	[550.15, 645.83]	[550.15, 645.83]	[−4126, −4844]	[0, 0]
0.4	[−936.9, −1056]	[562.11, 633.87]	[562.11, 633.87]	[−4216, −4754]	[0, 0]
0.6	[−956.8, −1037]	[574.07, 621.91]	[574.07, 621.91]	[−4306, −4664]	[0, 0]
0.8	[−976.7, −1017]	[586.03, 609.95]	[586.03, 609.95]	[−4395, −4575]	[0, 0]
1	[−996.7, −996.7]	[597.99, 597.99]	[597.99, 597.99]	[−4485, −4485]	[0, 0]
Unit time cost ranking	−1993.32	1195.98	1195.98	* −8969.9	0

Note: The * represented Project D attained a minimal ranking value.

.

Table 8. Defuzzification of each project based on α-cuts for unit time quality ranking values.

α	Project
	A	B	C	D	E
0	[−1.791, −2.189]	[−1.791, −2.189]	[−1.791, −2.189]	[−4.482, −5.478]	[0, 0]
0.2	[−2.189, −2.149]	[−2.189, −2.149]	[−2.189, −2.149]	[−4.582, −5.378]	[0, 0]
0.4	[−1.871, −2.109]	[−1.871, −2.109]	[−1.871, −2.109]	[−4.681, −5.279]	[0, 0]
0.6	[−1.91, −2.07]	[−1.91, −2.07]	[−1.91, −2.07]	[−4.781, −5.179]	[0, 0]
0.8	[−1.95, −2.03]	[−1.95, −2.03]	[−1.95, −2.03]	[−4.88, −5.08]	[0, 0]
1	[−1.99, −1.99]	[−1.99, −1.99]	[−1.99, −1.99]	[−4.98, −4.98]	[0, 0]
Unit time quality ranking	−3.98	−3.98	−3.98	* −9.96	0

Note: The * represented Project D, which exhibited a minimum value of −9.96 and was the lowest unit time quality.

represents the quality change rate per unit of time at various points. Project D signifies the lowest quality across all projects and falls below the 90.00% threshold. This minimal value will slightly decline over time and reduce Project D’s quality. These findings highlight the potential trade-off between time and quality, implying that enhanced quality necessitates increased costs.

The above analysis shows that time significantly affected Project D, resulting in an extended logistics delivery period. While the extension reduced costs, it also adversely affected the quality: Quality fell below the 90.00% standard. Therefore, this study chose to evaluate Project D’s impact on carbon emissions by FNMOPM; this study used Project D as an example for other projects. The result shows that extending the logistics delivery time in Project D could reduce carbon emissions from 5259.31 to 419,199.60 tons. The mathematical model representing the reduction of carbon emissions in Project D is as follows:

$$\begin{gathered} {∆C}_{FD}=\frac{c_{normal,FD}-c_{delay,FD}}{t_{delay,FD}-t_{normal,FD}}=\frac{\left(12557.86,13953.18, 15348.49\right)-(8521.40,9468.23,10415.05)}{\left(4.48,4.98,5.48\right)-(3.59, 3.99,4.39)} \\ =(68500,4500,1131.58) \end{gathered}$$

is the delay unit time cost in a fuzzy status.

$$\begin{gathered} a_{FD}=\frac{c_{normal,FD}-c_{delay,FD}}{t_{narmal,FD}^2-t_{delay,FD}^2}=\frac{\left(12557.86,13953.18, 15348.49\right)-(8521.40,9468.23,10415.05)}{{(3.59, 3.99,4.39)}^2-{\left(4.49,4.98,5.48\right)}^2} \\ =(-397.51,-501.68,-2423.84) \end{gathered}$$

is the fuzzy nonlinear time cost in curvature status.

$${CE}_{T,FD}=a_{FD}\times (t_{FD}{)}^2\times {(-CE}_{FD})+b_{FD}$$

$$\begin{gathered} {CE}_{T,FD}=\frac{c_{normal,FD}-c_{delay,FD}}{t_{normal,FD}^2-t_{delay,FD}^2}\times (t_{FD}{)}^2\times {(-CE}_{FD}) \\ +\frac{{c_{delay,FD}\times t}_{normal,FD}^2-c_{normal,FD}{\times t}_{delay,FD}^2}{t_{normal,FD}^2-t_{delay,FD}^2} \\ =\frac{\left(12557.86,13953.18, 15348.49\right)-(8521.40,9468.23,10415.05)}{{(3.59, 3.99,4.39)}^2-{\left(4.49,4.98,5.48\right)}^2}\times {\left(2.59, 2.99,3.39\right)}^2 \\ \times (2.15,2.38,2.63) \\ +\frac{(8521.40,9468.23,10415.05)\times {(3.59, 3.99,4.39)}^2-(12557.86,13953.18, 15348.49)\times {\left(4.49,4.98,5.48\right)}^2}{{(3.59, 3.99,4.39)}^2-{\left(4.49,4.98,5.48\right)}^2} \\ =\left(-2213.55,-3572.95,-21610.20\right)+\left(3045.77,21926.42,397589.39\right) \\ =\left(5259.31,25499.37,419199.60\right) \mathrm{tons} \end{gathered}$$

is the fuzzy nonlinear variation in Project D’s carbon emissions mitigation.

Based on $q_{normal,D}=85.00\mathrm{\%}$, $∆C_{Q_{normal,D}}=600 \mathrm{T}\mathrm{W}\mathrm{D}$ and $∆C_{Q_{delay,FD}}=3200 \mathrm{T}\mathrm{W}\mathrm{D}$, the fuzzy numerical value is developed as ${ q}_{normal,FD}=\left(76.24,84.72,93.19\right),∆C_{Q_{normal,FD}}=\left(538.19,597.99,657.79\right),\mathrm{a}\mathrm{n}\mathrm{d}∆C_{Q_{delay,FD}}=\left(2870.37,3189.30,3508.23\right)$. Project D exhibited the lowest unit time cost within the α-cut (interval based on the membership function). Project D was required to meet a 90.00% quality threshold. However, the mass ranges from −188.93% to 75.25%. Therefore, as Prajapati et al. discussed, companies should carefully consider the target market and service quality while making such decisions to avoid losing the customer base [40].

$$\begin{gathered} Q_{T,FD}=q_{delay,FD}+[\frac{{(q}_{delay,FD}-q_{normal,FD})}{(t_{delay,FD}-t_{normal,FD})}]\times \left(t_{FD}-t_{delay,FD}\right) \\ =\left(71.76,79.73,87.71\right)+[\frac{(\left(71.76,79.73,87.71\right)-\left(76.24,84.72,93.19\right))}{(\left(4.48,4.98,5.48\right)-\left(3.59,3.99,4.39\right))}]\times (\left(4.47,4.97,5.47\right) \\ -\left(4.48,4.98,5.48\right)) \\ =\left(71.76,79.73,87.71\right)+\left(115,-5,-11.32\right)\times \left(-1.01,-0.01,0.99\right) \\ =\left(-188.93,-0.99,75.25\mathrm{\%}\right) \end{gathered}$$

is a variation in the time quality of fuzzy Project D.

$$\begin{gathered} C_{Q,FD}=\left[(\frac{∆C_{Q_{delay,FD}-}∆C_{Q_{normal,FD}}}{t_{delay,FD}-t_{normal,FD}})\times \left(t_{FD}-t_{normal,FD}\right)+∆C_{Q_{delay,FD}}\right]\times [\left(q_{FD-}{Q}_{T,FD}\right)] \\ =[\left(\frac{\left(2870.37,3189.30,3508.23\right)-(538.19,597.99,657.79)}{(4.48,4.98,5.48)-(3.59,3.99,4.39)}\right)\times (\left(4.47,4.97,5.47)-(3.59,3.99,4.39\right))+ \\ (2870.37,3189.30,3508.23)]\times \left[\left(90,90,90\right)-(-188.93,-0.99,75.25)\right] \\ =\left(79586.17,522848.83,1591901.68\right)TWD \end{gathered}$$

is a variation in the quality cost of fuzzy Project D.

This study examined the potential for reducing carbon emissions within the logistics industry through extended delivery times, as explored in Project D. Project D demonstrated that extending delivery times could reduce carbon emissions by 5259.31 to 419,199.60 tons. However, this potential environmental benefit comes with a trade-off in delivery quality. Despite expanding the delivery windows, the quality would likely fall below the desired 90.00% standard and even below the original 85.00% expectation (ranging from −188.93% to 75.25%). Companies face a relative cost per unit quality ranging from TWD 79,586.17 to 1,591,901.68 to improve delivery quality. Consumers accept slightly longer delivery times, and consumers could benefit from reduced carbon emissions without incurring additional costs. Implementing a policy that encourages extended delivery times aligns with the long-term ESG objective of carbon emission mitigation by promoting sustainable practices within the logistics sector and effectively addressing carbon emissions. Project D highlights the potential for reducing carbon emissions in logistics through extended delivery times but emphasizes the critical trade-off with declining delivery quality. This information can inform policy decisions and consumer choices by providing insights into the sustainability–efficiency nexus within the logistics industry.

5. Discussion

The e-commerce industry faces increasing pressure to reduce the environmental impact of its logistics delivery systems. One potential solution is extending delivery times.

Project D revealed a crucial trade-off between sustainability and efficiency in e-commerce logistics delivery. While extending delivery times can significantly reduce carbon emissions, it can also lead to a decline in delivery quality, potentially affecting both businesses and consumers. This study is discussed below from several perspectives.

Balancing Efficiency and Sustainability: This article explores the trade-off between efficiency (time and cost) and sustainability (reduced carbon emissions) in e-commerce logistics. This research showed that quality dropped from 85.00% to 75.25%. To enhance the quality to 90.00%, companies must pay up to 1900% of the original price, 19 times the cost. Conversely, staying at 75.25% quality reduces payouts by 19 times. This is a payout ratio of 19:1. To remain at 75.25%, it only costs TWD 1, but to reach 90.00%, companies have to spend TWD 19 more. Consumers who reduce carbon emissions are willing to accept slower delivery speeds. Let e-commerce companies reduce carbon emissions costs and expenses and increase revenue. This brings optimistic prospects for the growth of the business.

Fuzzy Nonlinear Multi-Objective Programming Model (FNMOPM): This study proposes a novel FNMOPM to address complex decision-making problems in e-commerce logistics. Unlike traditional linear models used in previous research, this model considers fuzziness and nonlinearities in real-world scenarios. This study is grounded in optimization that closely mirrors actual situations. Enterprises should select a model that aligns with their specific context.

Time Management as a Strategy: This research suggests that time management can be a viable strategy for carbon emissions mitigation in e-commerce logistics. Extending delivery times while maintaining acceptable quality levels can be a solution. Although this study only performed illustrative calculations for Project D, it also provided an algorithm model. Other researchers can follow this algorithm model to calculate other projects.

Consumer Acceptance and Policy Implications: This article emphasizes the importance of consumer acceptance of longer delivery times to achieve significant emission reductions. Simultaneously, it encourages extended delivery windows to support sustainable development goals. Enterprises can provide two delivery modes for consumers: the standard delivery mode, with an arrival time of three days, or the carbon reduction delayed delivery mode, with an arrival time of five days.

6. Conclusions

6.1. Research Conclusions

This study focused on Project D, conducted by SASSWOOD Corporation in Taiwan. The current project aims to reduce carbon emissions in the logistics industry by extending delivery times and utilizing FNMOPM to address trade-offs related to TCQCE in e-commerce logistics. Due to its extended delivery window, Project E exhibited the highest potential for carbon emission reduction. Projects B and C emerged as cost-effective options. All of the above meet Hypothesis 1, reducing costs due to extended logistics delivery times.

This study used the quality of Project D as an example to calculate its cost and carbon emissions. Project D’s delivery time was adjusted, mitigating carbon emissions between 5259.31 and 419,199.60 tons. It is consistent with Hypothesis 2, which states that extended delivery times lead to lower carbon emissions.

The quality in Project D dropped from 85.00% to 75.25%. If the company wanted to improve its quality, the cost would increase between 79,586.17 and 1,591,901.68. It meets Hypothesis 3, stating that quality requirements will affect cost increases.

According to the above results, costs were 19 times higher, and carbon emissions were 78 times higher. The primary variable relationship was that this study used a nonlinear model as the algorithm, and that it was not a linear relationship. Nonlinearity is better for handling the variables of fluctuating situations. Thus, this study has successfully implemented an effective solution to reduce carbon emissions in the e-commerce logistics industry. Adopting a fuzzy nonlinear multi-objective programming model addresses gaps in previous research on extended delivery times.

Encouraging consumer acceptance of slightly longer delivery times can significantly reduce carbon emissions without compromising affordability. This study has made a successful contribution by taking advantage of FNMOPM to extend delivery times, which presents a promising time management approach for TCQEC.

Additionally, this study underscores the need for strategic interventions in the logistics delivery of the e-commerce industry. By extending delivery times while ensuring economic viability, companies can contribute to a greener future. As e-commerce evolves, consumers and governments must collaborate with companies on the ESG carbon emission mitigation goals to create a sustainable e-commerce logistics delivery industry, as discussed by Axsen, Plötz, and Wolinetz [41].

6.2. Research Recommendations

Future research should explore innovative solutions incorporating reverse or secondary logistics deliveries’ nonlinear cost and quality aspects. Since the e-commerce logistics operation model prioritizes accuracy and speed, integrating carbon reduction measures may impact efficiency and customer satisfaction. How can studies collectively create a more sustainable future? To mitigate the increase in carbon emissions and costs resulting from secondary logistics deliveries, one approach involves extending delivery times and utilizing FNMOPM to manage trade-offs. This effort requires consumers’ awareness of carbon reduction and effective time management. Encouraging collaboration among enterprises, consumers, and policymakers can create a win-win situation for environmental sustainability and efficient logistics operations.