Selecting Resilient Strategies for Cost Optimization in Prefabricated Building Supply Chains Based on the Non-Dominated Sorting Genetic Algorithm-Ⅱ: Facing Diverse Disruption Scenarios

Abstract

As a new sustainable building production mode, prefabricated building supply chains can realize energy saving, environmental protection and full cycle value maximization of building products. Prefabricated building supply chains often experience disruptions due to supply instability, transportation delay and force majeure, resulting in project delays and cost escalations and posing challenges to the sustainable development objectives of enterprises. Therefore, it is important and essential to study the strategy of enhancing the resiliency of prefabricated building supply chains, which has not been comprehensively explored in previous papers. This paper constructs decision-making models for supply chain cost resilience strategies under varying scenarios of supply disruptions, incorporating both redundant inventory and back-up supplier strategy. It considers the total cost and resilience of the supply chain as dual objective functions. Parameter-tuned non-dominated sorting genetic algorithm-Π (NSGA-Π) algorithms were used innovatively to solve the project case, and the impacts of the redundant inventory coefficient and back-up supplier supply price coefficient on the model result were analyzed. The results indicate that the supply chain with resilience construction has a superior capability to cope with disruption. The results show that when there is a mild supply disruption, the general contractor uses the capacity within the supply chain and chooses a redundant inventory strategy to restore resilience. In the event of moderate disruption, both the easy inventory strategy and back-up supplier strategy are selected to maintain supply chain stability. In the event of a severe disruption, only the back-up supplier strategy is selected to cover the losses and maintain the project schedule. In addition, the choice of resilience strategy is impacted by the inventory levels and component prices of back-up suppliers. It further verifies the effectiveness of the model and the impacts of uncertain parameters in the model on the results. This study contributes to enhancing the resilience management of the prefabricated building supply chain by the general contractor, thereby elevating the overall efficiency and competitiveness of the supply chain and furthering the sustainable development of prefabricated buildings.

1. Introduction

With the acceleration of globalization, the vulnerability of the supply chain has become increasingly prominent; it is susceptible to interference from both internal and external sources. In August 2015, biochemical explosions at the port of Tianjin in China went off without warning, destroying large stockpiles of cargo and causing the suspension of operations at two terminals. As a result, the supply chain was interrupted, causing huge losses [1]. In addition, the COVID-19 pandemic has drawn global attention to this matter, as nearly 97% of global supply chains were disrupted in a short span of time, leading to many companies suffering a loss of competitive edge and even permanent closure [2]. For example, in Wuhan, as an important city for the production of automobiles and their supporting products, almost all factories were closed during the COVID-19 period, and the supply of parts produced in China was interrupted, leading to the suspension of production by many multinational automobile manufacturing companies: Hyundai Motor’s factory in Korea stopped production partly because the parts purchased from China could not meet the demand; Mazda’s factories in Japan halted the production of some models because they could not source parts from China. Therefore, the ability of a company to maintain supply chain resilience after disruption is a key determinant of its competitive advantage and overall performance. It is necessary for the supply chain to possess the ability to recover from any disruption [3]. The resilience of the supply chain is one of the key factors to leverage in alleviating uncertainty and disruption, improving the efficiency of supply chain operation and promoting its sustainable development [4].

The construction industry is a major source of greenhouse gas emissions, and the construction supply chain involves the production, transportation and use of a large number of building materials and products, which are often accompanied by environmental pollution. However, the supply chain of the construction industry, unlike manufacturing, is more complex and unpredictable [5]. The large-scale construction of the construction industry has brought considerable economic benefits, but has also caused energy consumption and environmental pollution and other problems [6]. Strengthening the management of the construction supply chain can effectively promote the development of the construction industry to take on a more environmentally friendly, efficient and sustainable direction. As supply chains increase in complexity, enhancing resilience to mitigate the risk of sustainable disruption in the supply chain is an important issue [7]. Construction projects are typically one-off and customized and have long lead times. The characteristics of the construction supply chain make it susceptible to catastrophic events, such as COVID-19. Additionally, it is also vulnerable to daily operational risks, including traffic conditions and disruptions in material supply [8]. Construction supply chains are disrupted due to these uncertainties or risks, resulting in project delays or increased costs [9]. Therefore, strict risk management strategies are required to enhance the resilience of the construction supply chain. Additionally, flexible disruption response strategies are necessary. Supply chain disruptions are inherently unpredictable. When faced with uncertain supply disruptions, we can employ heuristic backward scenario reduction methods to generate random uncertain scenarios [10,11]. It is essential to consider all potential outage scenarios and formulate a flexible outage resilience strategy.

As a representative of green and low-carbon buildings, prefabricated buildings have been vigorously promoted by the state, with their advantages being efficient energy saving, emission reduction and construction efficiency. They can significantly improve production efficiency and reduce construction waste, which is considered to be an effective way for China’s construction industry to achieve the goal of double carbon, and prefabricated buildings have become the main trend and inevitable choice of green development. Compared with the traditional construction supply chain, the prefabricated construction project supply chain shows higher complexity in form diversity and cooperation dependence. Prefabricated construction improves building efficiency through standardized and modular production methods. It significantly reduces energy consumption and waste generation during on-site construction, contributes to the sustainable development of the construction industry and helps mitigate climate change. However, the production of prefabricated components and on-site assembly often occur concurrently, making the supply chain susceptible to disruptions due to changes in upstream information, which can lead to redundant work downstream [12]. Tight integration and real-time information exchange require high levels of coordination among all parties in the supply chain, including contractors, material suppliers, component manufacturers and construction firms [13]. Only through such collaboration can project progress be ensured and cost-effectively controlled. Therefore, the success of prefabricated construction projects depends largely on the stable operation of their supply chains. The major challenges in supply chain planning are mainly to maintain the stability of supply chains and improve their resilience to risks [14]. Enhancing supply chain resilience can effectively help them prevent potential risks and respond flexibly to potential disruptions.

The enhancement of supply chain resilience is still in its developmental stage within the field of prefabricated construction [15]. Recently, considerable research has focused on identifying the influencing factors of supply chain resilience and the interactions between these factors. However, there is limited research on how supply chains recover levels of resilience after disruptions. Based on this, this paper establishes a supply chain cost resilience strategy decision model. The model includes all possible disruption scenarios and takes the minimization of the total supply chain cost and the maximization of resilience as double objective functions, considering two resilience strategies: redundant inventory and back-up supplier. Parameter-tuned NSGA-Π algorithms were used innovatively to solve the case, and the optimal combination of resilience strategies after supply chain disruption was found, verifying the effectiveness of the model.

2. Literature Review

At present, the study of supply chain resilience focuses on two primary aspects. Firstly, there is a focus on accurately measuring the resilience of supply chains. Wang et al. used an interpretive structural model and an ANP-Fuzzy comprehensive evaluation model to measure the resilience of a green building supply chain [16]. Joshi et al. used the SWARA application to evaluate supply chain resilience innovation [17]. Belhadi et al. proposed a method that combines recovery time and financial performance indicators to more accurately quantify the resilience of the supply chain [18]. Fu et al. adopted a multi-dimensional competitiveness assessment framework to provide a powerful tool for a comprehensive assessment of supply chain resilience [19]. Secondly, attention is directed towards exploring effective strategies for enhancing supply chain resilience. Zhang et al. studied three scenarios in which manufacturers’ emission reduction investment strategies in the supply chain equalize the economic environment [20]. Vanessa Klementzki et al. compared some of the concepts involved in supply chains and listed proactive measures to improve supply chain resilience after disruptions [21]. Gong et al. constructed a two-stage adaptive mixed-integer fraction programming model to optimize supply chain resilience and economic cost and integrate network configuration, equipment capacity and capital cost [22]. Additionally, Hsu et al. built a decision framework that integrates a two-stage House of Quality and multicriteria decision-making approach, which aims to improve supply chain resilience and reduce sustainable supply chain risks through big data analytics [23]. Sonia Irshad Mari et al. developed quantitative resilience standards for supplier selection and order allocation to enhance supply chain resilience in a fuzzy environment [24]. Yan et al. compared the strategies adopted by each participant in the supply chain to determine the optimal investment strategy for supply chain recovery and supply capacity enhancement [25]. Furthermore, in terms of research methods for supply chain resilience, Rehman et al. employed a multi-criteria decision-making approach to study resilience recovery strategies in healthcare supply chains [26]. Mohammad Fattahi et al. used the supply network method of a two-stage stochastic program to study the recovery time and cost of a supply chain after disruption events [27]. Li et al. adopted an improved genetic algorithm to study the supply chain hypernetwork resilience optimization strategy based on product family [28].

However, the supply chain of a prefabricated construction project is not a simple industry chain or process, but a complex network of relationships composed of multiple stages and disciplines. Prefabricated construction projects typically involve numerous participants and complex processes of design, production, transportation, assembly and construction. There are many reasons for disruptions in the prefabrication supply chain, such as a shortage of construction workers, material supply disruptions, order backlogs and delays in the delivery of prefabricated components. Additionally, the materials and components in the supply chain for prefabricated construction projects are typically large in size and heavy, requiring on-site processing and assembly. This complexity in logistics and transportation leads to supply disruptions, potentially delaying project progress and increasing project costs. To reduce the adverse effects of supply chain disruptions, maintaining high supply chain resilience is essential; this proactive approach plays a pivotal role in ensuring the stable operation of the supply chain for prefabricated construction projects [29].

Strategies to enhance supply chain resilience include selecting the right suppliers, managing inventory redundancies and determining transportation modes and routes. Keskin et al. proposed a simulation optimization approach to address the issues of supplier selection and inventory management, incorporating stochastic demand and supply disruptions to enhance resilience [30]. Rajesh and Ravi proposed a gray relational analysis method for selecting back-up suppliers, which is based on the weights of supplier attributes, thereby enhancing the resilience of the supply chain [31]. Namdar et al. calculated the cost of starting back-up suppliers during supply chain disruptions using a hybrid multi-criteria decision-making approach that takes into account the four resilience dimensions of anticipation, preparedness, robustness and recovery [32]. Zhang et al. established a cost-resilience model to enhance supply chain resilience, using the MOPSO method to determine the optimal transportation modes and routes for prefabricated components [33].

At present, the research on prefabricated building supply chain mainly focuses on the production scheduling of components. Maziar Yazdani used a genetic algorithm, differential evolution and an imperialist competitive algorithm to study the production scheduling problem of prefabricated components [34]. Yang et al. proposed a scheduling optimization model for prefabrication that focuses on multiple production lines [35]. Wang and Hu analyzed the impact of mold manufacturing, prefabricated component storage and transportation on the supply chain and improved the process shop scheduling model from a supply chain perspective [36]. Xie et al. constructed a production scheduling model based on the just-in-time strategy [37]. Wang et al. summarized the hot spots and gaps in the research on component production scheduling by combing the literature [38]. However, their work is limited to the study of the supply chain resilience of prefabricated construction projects. The current research mainly focuses on the influencing factors. Zhang et al. utilized the fuzzy set theory, decision experiment analysis and system dynamics to investigate the influencing factors of supply chain resilience in prefabricated construction [29]. Hua et al. determined the key influencing factors of the supply chain resilience of prefabricated buildings and the impact of BIM on the key factors by adopting the method of comprehensive review [39]. Zhang et al. used a structural equation model to analyze the impacts of various factors on the supply chain resilience of prefabricated buildings. Their study concluded that the production of components and the assembly construction process significantly affect supply chain resilience [33].

In the field of prefabricated building supply chains, there has been little research on strategies that can simultaneously improve resilience and control costs after supply disruptions. Some effective resilience strategies inevitably lead to additional costs, such as back-up suppliers, and redundant inventory strategies. Therefore, it is crucial to balance resilience and cost when enhancing the supply chain resilience of prefabricated buildings. [40]. When considering the redundant inventory strategy, whether the strategy is chosen depends largely on the amount of redundant inventory. This is because the higher the amount of redundant inventory, the greater the inventory management costs. Similarly, for the alternative supplier strategy, whether the strategy is selected is mainly determined by the supply price of the back-up supplier.

The conflict between supply chain resilience and total cost often presents a classic multi-objective optimization problem. There are many algorithms to solve multi-objective optimization problems, such as the particle swarm optimization algorithm, ant colony optimization algorithm, genetic algorithm, NSGA-Π and NSGA-III. Compared to other algorithms, the NSGA-Π converges faster and has higher solution accuracy when seeking Pareto front solutions. It can efficiently find Pareto front solutions through non-dominant sorting and congestion comparison mechanisms. Additionally, the NSGA-Π has high flexibility and tunability. The algorithm parameters can be flexibly adjusted according to the characteristics of the problem to achieve improved optimization results. It can provide decision makers with a comprehensive solution. Therefore, the NSGA-Π is widely used as a multi-objective optimization algorithm in reality [41], especially in solving double-objective optimization problems [42].

Supply chain resilience is considered a relatively new concept in the prefabricated construction industry. Current research focuses on identifying the influencing factors and their interactions in the prefabricated building supply chain. There are still few studies on how to select strategies to enhance the resilience of prefabricated building supply chains after supply disruptions. Therefore, this paper constructs a cost resilience strategy decision model of a prefabricated building supply chain after supply disruption, which considers the redundant inventory strategy and back-up supplier strategy and regards the total cost and resilience of the supply chain as dual objective functions. The NSGA-Π is proposed to solve the problem, and the Pareto solution set is analyzed. It is found that the supply chain with resilience construction has a better ability to cope with supply disruption, which can reduce the total cost of the supply chain after supply disruption to a certain extent, verifying the effectiveness of the model.

3. Methods

3.1. Supply Disruption Scenarios

Suppliers may fail to deliver prefabricated components on time due to various issues, such as natural disasters, equipment failures, raw material shortages or labor problems. This is particularly critical for high-demand items like precast slabs, columns and beams. During the transportation process, there may be unexpected situations such as traffic jams or traffic accidents. In addition, prefabricated components, due to their large size and heavy weight, may be damaged during transportation and need to be returned to the factory for repair. This can result in prefabricated components not arriving at the construction site on time. This is especially true for prefabricated components with complex shapes, such as precast stairs and precast balcony slabs.

As shown in Figure 1, a three-level supply chain consisting of the production, transportation and installation of prefabricated building components is studied. In large prefabricated construction projects, due to the high demand for prefabricated components, multiple component suppliers and back-up suppliers are selected by the general contractor. The quantity of prefabricated components supplied by each supplier varies greatly, resulting in the number of components lost being out of stock due to supply disruptions, which means that different types of components have different levels of disruption in different scenarios. Therefore, different suppliers’ supply disruptions and different types of prefabricated component disruptions are considered when analyzing supply disruption scenarios.

3.2. Mathematical Model

In the event of supply disruption, it is a complex issue to choose resilience strategies to maintain the economics and stability of the supply chain. As shown in Figure 2, the resilience strategy selection of the construction general contractor is studied through the following steps: The main purpose of step 1 is to describe the disruption scenarios that occur more frequently. The main purpose of step 2 is to formulate hypotheses and establish a mathematical model. It is based on various scenarios of supply disruptions, with the objective functions being the total supply chain cost and supply chain resilience, taking into account strategies such as redundant inventory and back-up suppliers. The main purpose of step 3 is to obtain the optimal parameter combination that can make the NSGA-Ⅱ run efficiently by analyzing the two parameters of PC and PM in the NSGA-Ⅱ. The main purpose of step 4 is to run the tuned NSGA-Ⅱ using Python to solve the mathematical model. The conclusion is drawn by analyzing the result of the algorithm’s operation. The main purpose of step 5 is to analyze the impact of the redundant inventory coefficient and back-up supplier supply price coefficient on the selection of the resilience strategy. It is beneficial to improve the decision-making level of the construction general contractor in different prefabricated construction projects.

3.2.1. Assumptions

To enhance supply chain resilience, we will optimize strategic choices in the prefabricated building supply chain after supply disruptions. The recovery conditions of the supply chain are restricted after the occurrence of the disruption event, and the model is constructed on the basis of the following assumptions:

• Supply disruptions occur during the ordering period, and the contractor’s redundant inventory of prefabricated components is not considered;
• The back-up supplier strategy is initiated only when the supply is interrupted, and the supply cost of the back-up supplier is higher than that of the on-chain supplier;
• The replenishment time of the component whose supply is disrupted is less than the allowable replenishment time;
• The unit supply costs for component suppliers, redundant start-up costs, inventory management costs, installation costs for construction contractors and unit out-of-stock costs in the event of supply disruptions are known.

3.2.2. Objective Function

The total cost of the prefabricated building supply chain $C$ is mainly obtained by adding up the cost of running a supply chain after supply disruption $C_1$, the cost of choosing a back-up supplier strategy after supply disruption $C_2$, the cost of choosing a redundant inventory strategy after supply disruption $C_3$ and the shortage cost $C_4$. $C_1$ is composed of the supply cost, transportation cost and installation cost of the supplier’s normal supply components; $C_2$ is composed of the supplier’s redundant supply cost, transportation cost and installation cost; $C_3$ is composed of the supply cost, Stack management cost, transportation cost and installation cost of the prefabricated components supplied by the back-up suppliers; and $C_4$ is a component of the out-of-stock cost caused by the actual supply quantity of components not meeting the demand of the construction unit. The details can be determined according to Equations (1)–(5):

$$Min{f}_1=MinC={\displaystyle \sum _{n\in N}{P}_n(C_1+C_2+C_3+C_4)}$$

(1)

where $P_n$: the probability of scenario $n$ occurring.

$$C_1=(P{C}_{ij}+T{C}_{ij}+S{C}_{ij})Q_{nij}$$

(2)

where $C_1$: the cost of running the supply chain after supply disruption;

$P{C}_{ij}$: the unit cost of prefabricated component $j$ supplied by supplier $i$ to the construction contractor;

$T{C}_{ij}$: the unit transport cost of the prefabricated component $j$ from supplier $i$ to the construction contractor;

$S{C}_j$: the unit installation cost of prefabricated component $j$;

$Q_{nij}$: the quantity of prefabricated component $j$ supplied by supplier $i$ to the construction contractor in disruption scenario $n$.

$$C_2=Z_{1n}{\displaystyle \sum _{k=1}^{K}P{C}_k^0+{\displaystyle \sum _{k=1}^K{\displaystyle \sum _{j=1}^J(PC\times \omega +T{C}_{kj}+S{C}_j)Q_{nkj}}}}$$

(3)

where $C_2$: the cost of choosing the back-up supplier strategy after supply disruption;

$Z_{1n}$: represents a value of 1 if the strategy of the back-up supplier is adopted and otherwise has a value of 0 in disruption scenario $n$;

$P{C}_k^0$: the start-up costs of choosing back-up supplier $k$;

${\omega }_k$: the supply price coefficient of back-up supplier $k$;

$T{C}_{kj}$: the unit transport cost of prefabricated component $j$ from back-up supplier $k$ to the construction contractor;

$Q_{nkj}$: the quantity of prefabricated component $j$ supplied by back-up supplier $k$ to the construction contractor in disruption scenario $n$.

$$C_3=Z_{2n}{\displaystyle \sum _{i=1}^{I}F{C}_i^0+{\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^J({FC}_{ij}+{TC}_{ij}+{MC}_{ij}+{SC}_j)F{Q}_{nij}}}}$$

(4)

where $C_3$: the cost of choosing a redundant inventory strategy after supply disruption;

$Z_{2n}$: represents a value of 1 if the strategy of redundant inventory is adopted, and otherwise, it is 0 in disruption scenario $n$;

$F{C}_i^0$: the redundant inventory’s fixed cost of supplier $i$;

$F{C}_{ij}$: the redundant supply cost of prefabricated component $j$ provided by supplier $i$ to the construction contractor;

$M{C}_{ij}$: the inventory management cost of prefabricated component $j$ at supplier $i$;

$F{Q}_{nij}$: the redundant supply quantity of prefabricated component $j$ supplied by supplier $i$ to the construction contractor in disruption scenario $n$.

$$C_4=SOC({\displaystyle \sum _{j=1}^J{D}_j-({\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^J(Q_{nij}+F{Q}_{nij)}+{\displaystyle \sum _{k=1}^K{\displaystyle \sum _{j=1}^J{Q}_{nkj}))}}}}}$$

(5)

where $C_4$: the shortage cost;

$SO{C}_j$: the unit out-of-stock cost of prefabricated component $j$;

$D_j$: the demand of the construction contractor for prefabricated component $j$.

For prefabricated building supply chains, the primary goal is to plan and direct the necessary quantity of materials to the site for final assembly [43]. Therefore, the quantification of the resilience level for such supply chains should consider the discrepancy between the supply quantity and actual demand after a supply interruption. The details can be determined according to Equation (6):

$$Max{f}_2=Max\gamma =({\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^J(Q_{nij}+F{Q}_{nij})+{\displaystyle \sum _{k=1}^K{\displaystyle \sum _{j=1}^J{Q}_{nkj})/{\displaystyle \sum _{j=1}^J{D}_j}}}}}$$

(6)

where $\gamma$: supply chain resilience.

3.2.3. Constraint Condition

Depending on the problem being studied, the constraints are listed as follows:

$${\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^J{Q}_{nij}}}\le (1-X_{ni})\times Gca{p}_i+X_{ni}\times {\theta }_{ni}\times Gca{p}_i$$

(7)

$${\displaystyle \sum _{j=1}^J{D}_j-({\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^J{Q}_{nij}+F{Q}_{nij})-{\displaystyle \sum _{k=1}^K{\displaystyle \sum _{j=1}^J{Q}_{nkj}}\ge 0}}}}$$

(8)

$$0\le {\displaystyle \sum _{i=1}^I{\displaystyle \sum _{j=1}^{J}F{Q}_{nij}\le Z_{2n}\times X_{ni}\times {\mu }_i\times Gcap}}$$

(9)

$$0\le {\displaystyle \sum _{k=1}^K{\displaystyle \sum _{j=1}^J{Q}_{nkj}\le Gca{p}_k\times Z_{1n}}}$$

(10)

$$Z_{1n},Z_{2n}\in \left\{0,\right.\left.1\right\}$$

(11)

where $X_{ni}$: represents a value of 1 in disruption scenario $n$ if supplier $i$ experiences a supply disruption, and otherwise, it is 0;

$Gca{p}_i$: the maximum production capacity of supplier $i$ (excluding the redundant inventory capacity);

${\theta }_{ni}$: the capacity ratio of failed suppliers under disruption scenario $n$;

${\mu }_i$: the redundant inventory capacity coefficient of supplier $i$;

$Gca{p}_k$: the maximum production capacity of back-up supplier $k$.

Constraint (7) outlines that the normal supply quantity of the supplier on the chain cannot exceed the maximum supply capacity of the supplier in varying supply disruption scenarios. Constraint (8) outlines that the total of the normal supply quantity of the on-chain supplier and the supply quantity derived from either the back-up supplier strategy or the redundant inventory strategy shall not exceed the demand of the construction in different supply disruption scenarios. Constraint (9) outlines that the quantity supplied cannot exceed the quantity of redundant inventory when adopting the redundant inventory strategy. Constraint (10) outlines that the quantity supplied cannot exceed the maximum supply capacity of the back-up supplier when selecting the back-up supplier strategy. Constraint (11) determines whether the back-up supplier strategy or the redundant inventory strategy is to be selected.

3.3. NSGA-Π

The NSGA-Π was proposed by Srinivas and Deb in 1995 [44]. It is a multi-objective optimization algorithm based on the principle of the genetic algorithm. It seeks Pareto frontier solutions through non-dominant sorting and congestion allocation strategies. In the NSGA-II, initial populations are first randomly generated, and each individual is assessed for fitness. Subsequently, the individuals are ranked according to their non-dominant relationships, and the crowding degree of the individuals within each rank is calculated to maintain the diversity of the population. In the selection operation, the tournament selection algorithm is used to compete according to the rank and crowding of individuals, and the better individuals are selected for genetic operations, including crossover and mutation, to produce new individuals. Finally, the newly generated individuals replace some of the original individuals, forming the next generation of populations. Through continuous iteration, the NSGA-Π can search for a set of Pareto frontier solutions with excellent performance on multiple objective functions in multi-objective problems and provide multiple options for decision makers.

As is well known, the chromosome is the basic unit in the NSGA-Π and represents a solution in the solution space. The chromosome length refers to the number of genes in the chromosome, which is directly related to the complexity of the problem and the way it is represented. In the supply chain cost resilience strategy decision model, the integer coding of the chromosome is used, the chromosome length indicates the number of decision variables, which determines whether or not a resilience strategy should be selected and the quantity of various components supplied by normal suppliers along with the selected resilience strategies.

4. Case Study

4.1. Project Description

The China Overseas Real Estate Dangjia Smart City project is taken as a case study, which is located in Jinan, Shandong Province, China. The project consists of 10 buildings. Buildings 1#, 2#, 3#, 5#, 6# and 7# all have 18 stories above ground and 2 floors below ground; Buildings 8# and 9# have 17 stories above ground and 2 floors below; Building 10# has 16 stories above ground and 2 floors below and Building 11# has 15 stories above ground and 2 floors below. The total construction area is 136,652.46 m², and the overall assembly rate is 50%. It has three component suppliers and one back-up supplier. In the prefabricated component supply chain of the prefabricated project, in a class of components where supply disruptions are easily caused and demand is high, the prefabricated board is chosen for analysis. In a class of components whose shapes are complex and difficult to transport, prefabricated stairs are selected for analysis. As shown in

Table 1. Basic parameters of suppliers and construction units.

Supplier	Component Type	Unit Supply Cost	Administrative Cost Unit of Redundancy	Unit Redundant Supply Cost	Redundant Inventory Start-Up Costs	Maximum Productivity
I1	Precast slab	2550	400	2570	28,000	700
	Precast stairs	2520	520	2550		60
I2	Precast slab	2580	450	2600	30,000	500
	Precast stairs	2480	500	2500		80
I3	Precast slab	2630	380	2650	31,000	800
	Precast stairs	2550	500	2580		70

,

Table 2. Basic parameters of prefabricated components.

	Installation Cost	Quantity Demanded	Unit Out-of-Stock Cost
J1 (Precast slab)	34	1900
J2 (Precast stairs)	292	200

and

Table 3. Basic parameters of back-up supplier.

Supplier	Start-Up Cost	Unit Supply Cost of Precast Slab	Unit Supply Cost of Precast Stairs	Maximum Productivity of Precast Slab	Maximum Productivity of Precast Stairs
K1	38,000	2845	2768	800	60

, the market price of prefabricated components, the demand for components and the installation costs and inventory management costs can be determined using the survey data. The supplier’s information indicates that for major assembly projects, there are one to two levels of demand for redundant inventory. Therefore, a redundant inventory coefficient is introduced, which is set at 10%. The back-up supplier’s supply price is generally more expensive than the supplier’s supply price in the supply chain. Based on the actual situation of the China Overseas Real Estate Dangjia Smart City project, we introduced a back-up supplier unit supply price coefficient and set it at 1.1. The production of prefabricated components is a continuous process that requires close coordination of all links. Once a link is interrupted, the entire production process is affected. Supplier outages can lead to a shortage of raw materials and the inability of production lines to operate normally, which can significantly reduce capacity ratios. In the project, in order to flexibly respond to different scenarios of supply disruption, the surplus capacity ratio is set at 60%.

In real life, the possibility of three suppliers having a supply disruption at the same time is relatively small; this scenario is not discussed. As shown in

Table 4. Classification of supply disruption scenarios.

	Scenario 1	Scenario 2	Scenario 3	Scenario 4
Supply disruption	I1	I2	I1, I3	I1, I2
Probability	0.3	0.3	0.2	0.2

, different disruption scenarios are analyzed. Suppliers 1, 2 and 3 have different supply capacities for prefabricated components, and the damage quantity of components is different due to the supply interruption. In addition, the damage quantity of components caused by the supply interruption of one supplier and the simultaneous interruption of two suppliers is also different, which means that the prefabricated components will have different degrees of damage. First, the supply disruption scenario for supplier 1 is analyzed. The supplier’s ability to supply each prefabricated component is different, and the general contractor’s order quantity of components from each supplier is also different, which leads to different levels of disruption for different prefabricated components. Therefore, it is necessary to analyze different suppliers’ supply disruption scenarios. Since supplier 2 and supplier 3 have similar supply capabilities, we only analyze supply disruption scenarios for supplier 1 or supplier 2. In scenario 1, supplier 1 has a supply disruption. The out-of-stock rate of the precast slab is greater than that of precast stairs. In scenario 2, supplier 2 has a supply disruption. The out-of-stock rate of precast stairs is greater than that of the precast slab. Similarly, when the supply of two suppliers is disrupted, it is also necessary to consider the supplier’s supply capacity and the general contractor’s order volume. In addition, the severity of the disruption relative to one supplier needs to be considered. In scenario 3, supplier 1 and supplier 3 experience a supply disruption. The out-of-stock rate of the precast slab is greater than that of precast stairs, but the degree of disruption is more serious than that in scenario 1. In scenario 4, supplier 1 and supplier 2 experience a supply disruption. The out-of-stock rate of precast stairs is greater than that of the precast slab, but the degree of disruption is more severe than that in scenario 2. These two disruption scenarios already cover the issues we want to consider, so the disruption scenarios of both supplier 2 and supplier 3 are not analyzed.

4.2. NSGA-II Parameter Tuning Analysis

The results of the NSGA-II are affected by four adjustable parameters: the population number, N; the crossover probability, PC; the mutation probability, PM; and the number of iterations, T. Among them, if T is too small, there will obviously be inbreeding, producing sick genes. If T is too large, the result will be difficult to converge and waste resources, and robustness will decline. According to previous studies in the literature, the range of T is generally (40,1000), and since the range of our solution space is relatively large, we set T to 500 [45]. In order to stabilize the solution results, the number of iterations was set to 150. Crossover refers to the process that the structure of two parent individuals exchange with each other to produce new individuals according to a certain probability, PC. Variation is the process of randomly changing an individual’s gene value from “0” to “1” or from “1” to “0” with a small probability, PM. The sizes of PC and PM have a great impact on the optimization efficiency of the algorithm. The larger the PC is set, the lower the proportion of chromosome structure from the current parent copied to the next generation, and the higher the proportion of chromosome structure exchanged with another parent, resulting in more damage to the parent’s chromosome structure. Conversely, setting a smaller PC leads to less efficient updating of the chromosome structure. A larger PM setting implies a greater likelihood of mutation in the parent chromosome during subsequent iterations of optimization; conversely, a smaller PM setting makes it easier for effective genes to be rapidly lost and difficult to repair. Therefore, in order to enhance the algorithm’s optimization efficiency, it is essential to determine an optimal value combination for PC and PM through evaluation indices.

In the model, supply chain cost and resilience are two objective functions, and it is necessary to evaluate the combination of PC and PM by comparing the values of the two objective functions. In addition, the hypervolume index (HV) refers to the volume of the region in the object space surrounded by the non-dominated solution set and the reference point obtained by the algorithm. The larger the HV value, the better the comprehensive performance of the algorithm. The running time represents the time it takes for the algorithm’s results to stabilize. The shorter the running time, the better the algorithm’s performance. Therefore, two pre-Pareto evaluation metrics, the HV and runtime, are used to evaluate the performance and efficiency of the algorithm to find the optimal combination of PC and PM values. The setting ranges of the crossover probability and mutation probability are from 0.4 to 0.8 and from 0.02 to 0.1. The intervals for adjusting these parameters are 0.1 and 0.02, respectively, resulting in a comprehensive set of 25 experimental groups [46].

As shown in Figure 3a,b, the two objective function values have no regularity in different combinations of PC and PM values. Therefore, it is not feasible to evaluate parameter combinations with only two objective function values. It is necessary to introduce the HV and running time index to evaluate the optimal combination of parameters.

Table 5. The values of the two objective functions and two valuation indexes of the Pareto solution set under different parameter combinations.

PC	PM	C	$\mathit{\gamma }$	HV	Runtime
A1 = 0.4	B1 = 0.02	5,619,212–5,672,990	0.936286–0.985857	0.438499	45.492999
	B2 = 0.04	5,612,002–5,679,952	0.936762–0.992000	0.761009	44.618267
	B3 = 0.06	5,615,478–5,678,081	0.941667–0.992381	0.470446	45.401857
	B4 = 0.08	5,615,478–5,682,618	0.948857–0.993667	0.459055	45.303463
	B5 = 0.1	5,614,852–5,673,872	0.896905–0.995857	0.915225	54.792481
A2 = 0.5	B1 = 0.02	5,620,432–5,677,372	0.935714–0.976905	0.771423	54.934436
	B2 = 0.04	5,615,659–5,700,653	0.944667–0.989	0.558138	45.098157
	B3 = 0.06	5,612,891–5,668855	0.937810–0.984619	0.859872	41.611314
	B4 = 0.08	5,611,186–5,701,643	0.940857–0.993333	0.6777943	41.654660
	B5 = 0.1	5,614,117–5,692,019	0.941–0.995714	0.730285	41.586862
A3 = 0.6	B1 = 0.02	5,618,831–5,707,194	0.944381–0.991333	0.583920	42.260226
	B2 = 0.04	5,614,311–5,729,062	0.937762–0.991238	0.638460	41.921171
	B3 = 0.06	5,613,660–5,672,558	0.940905–0.990905	0.856819	38.474771
	B4 = 0.08	5,618,351–5,705,898	0.947143–0.993381	0.529488	41.715803
	B5 = 0.1	5,612,838–5,692,943	0.940905–0.991	0.423849	42.094688
A4 = 0.7	B1 = 0.02	5,615,586–5,674,965	0.938143–0.982048	0.881516	39.001013
	B2 = 0.04	5,615,495–5,671,597	0.948762–0.985857	0.337527	39.616937
	B3 = 0.06	5,616,655–5,680,835	0.942238–0.976905	0.743401	40.141561
	B4 = 0.08	5,619,062–5,716,382	0.938857–0.991905	0.605097	40.366962
	B5 = 0.1	5,608,083–5,700,971	0.944667–0.993286	0.343384	40.562592
A5 = 0.8	B1 = 0.02	5,620467–5,694,035	0.939571–0.988429	0.869571	38.631608
	B2 = 0.04	5,614,499–5,680,056	0.940333–0.993667	0.365646	38.475209
	B3 = 0.06	5,614,860–5,676,519	0.938190–0.992476	0.391121	47.790293
	B4 = 0.08	5,615,052–5,673,303	0.941762–0.983429	0.396022	45.733700
	B5 = 0.1	5,619,659–5,765,266	0.943857–0.988714	0.799799	36.526029

shows the results under 25 different parameter combinations. The HV values range from 0.33 to 0.92, and the running time ranges from 36 s to 55 s. The maximum running time is 54.934436 s, which is relatively shorter and within the acceptable range. Therefore, the HV can be given priority when evaluating the combination of PC and PM parameters. Six groups of parameter combinations, A1B5, A2B3, A3B3, A4B1, A5B1 and A5B5, were initially selected as the optimal parameter combinations. As shown in Figure 4, we analyzed the average of the two objective function values. The total average of the total supply chain costs was 5,563,167, and the total average of supply chain resilience was 0.964438. The two objective functions studied in this paper are the minimum total cost and maximum resilience of the supply chain. Therefore, the total average cost of the supply chain under the combination of A5B1 and A5B5 is higher than the total average, which fails to satisfy the requirements of the optimal parameter. The average resilience of the supply chain under the combinations of A1B5, A2B3 and A4B1 is lower than the total average resilience, which does not meet the requirements of the optimal parameter. Under the A3B3 combination, the average total cost of the supply chain is lower than the average cost, and the average resilience is higher than the average resilience, which satisfies the requirements of the optimal parameter combination. In summary, A3B3 is the optimal parameter grouping; PC = 0.6 and PM = 0.06. Figure 5 shows the Pareto optimal solution set under the optimal parameter combination.

4.3. Optimization Results

Figure 5 shows the trade-off between prefabricated building supply chain resilience and total cost. The results indicate that the improvement in supply chain resilience leads to the increase in the total supply chain cost.

As shown in

Table 6. Comparison of total supply chain cost and resilience in three scenarios.

	Total Supply Chain Cost	Supply Chain Resilience
Normal operation	5,626,120	1
Supply disruption	5,883,680	0.6
Resilience management	5,653,167	0.964438

, the total cost of the supply chain under normal circumstances is CNY 5,626,120, and the toughness level is 1. The total cost of the supply chain without resilience optimization after supply interruption is CNY 5,883,680, and the resilience level is 0.6. In the case of supply interruption, the combined resilience strategy of redundant supply and the alternative supplier is adopted to minimize the total cost and maximize resilience, and the cost after re-optimization is CNY 5,653,167, and the resilience level is 0.964438 yuan. By comparing the cost of the supply chain in the three situations, it can be found that when the first batch of supply interruption occurs, if there is no resilience management, the cost will increase greatly, and the resilience level will be greatly reduced. However, the adoption of a resilience strategy in the supply chain can reduce the cost after supply interruption to a certain extent and enhance the resilience level of the supply chain. The cost did not increase significantly, which also verified the effectiveness of the cost resilience model of a prefabricated building supply chain.

Construction general contractors can flexibly select the optimal solution from Pareto front solution sets based on actual needs. It can achieve targeted optimization under supply disruption scenarios by balancing resilience and cost. For example, if the construction general contractor pays more attention to the total cost of the supply chain rather than resilience, the scheme with a relatively low resilience level but the lowest total cost is selected, which corresponds to the leftmost solution of the Pareto front solution set shown in Figure 5.

Table 7. Optimal solutions for different scenarios from a cost minimization perspective.

Supplier			Scenario 1	Scenario 2	Scenario 3	Scenario 4
I1	Normal supply	Precast slab	419	698	372	417
		Precast stairs	33	51	33	31
	Redundant supply	Whether or not to adopt	Yes	Yes	No	Yes
		Precast slab	101	100	—	107
		Precast stairs	7	6	—	6
I2	Normal supply	Precast slab	499	299	485	299
		Precast stairs	78	44	72	47
	Redundant supply	Whether or not to adopt	No	No	No	No
		Precast slab	—	—	—	—
		Precast stairs	—	—	—	—
I3	Normal supply	Precast slab	798	799	458	797
		Precast stairs	68	68	37	0
	Redundant supply	Whether or not to adopt	No	No	No	No
		Precast slab	—	—	—	—
		Precast stairs	—	—	—	—
back-up supplier	Whether or not to adopt		No	No	Yes	Yes
	Precast slab		—	—	581	570
	Precast stairs		—	—	50	57

shows the specific solution. In scenarios 1 and 2, redundant inventory strategy I₁ is chosen to respond to disruptions and enhance resilience. In scenario 3, back-up supplier strategies are selected to make up for the shortage. In scenario 4, two strategies adopting redundant inventory of I₁ and back-up suppliers are chosen to maintain the normal operation of the supply chain.

If the construction general contractor pays more attention to the resilience level of the supply chain rather than the total cost, the scheme with a relatively high resilience level but the highest total cost is selected, which corresponds to the rightmost solution of the Pareto front solution set shown in Figure 5.

Table 8. Optimal solutions for different scenarios under the resilience maximization perspective.

Supplier			Scenario 1	Scenario 2	Scenario 3	Scenario 4
I1	Normal supply	Precast slab	417	696	373	416
		Precast stairs	33	52	32	31
	Redundant supply	Whether or not to adopt	Yes	Yes	Yes	Yes
		Precast slab	102	103	—	—
		Precast stairs	6	6	—	—
I2	Normal supply	Precast slab	498	297	487	298
		Precast stairs	78	44	72	46
	Redundant supply	Whether or not to adopt	Yes	No	No	No
		Precast slab	71	—	—	—
		Precast stairs	8	—	—	—
I3	Normal supply	Precast slab	797	797	457	793
		Precast stairs	68	68	39	67
	Redundant supply	Whether or not to adopt	No	No	No	No
		Precast slab	—	—	—	—
		Precast stairs	—	—	—	—
back-up supplier	Whether or not to adopt		No	No	Yes	Yes
	Precast slab		—	—	581	391
	Precast stairs		—	—	55	55

shows the specific solution. In scenario 1, redundant inventory strategies I₁ and I₂ are selected simultaneously to deal with disruption. In scenario 2, redundant inventory strategy I₁ is chosen to enhance resilience. In scenario 3 and scenario 4, back-up supplier strategies are selected to make up for the shortage and maintain the stability of the supply chain.

This provides decision makers with the flexibility to choose the best strategy based on different priorities and objectives. We can conclude that its resilience strategy varies with the degree of disruption. Supply chains rely primarily on internal capabilities, and a redundant inventory strategy is selected to control the adverse consequences of disruptions when there are minor supply disruptions. Both the internal redundancy inventory strategy and external back-up supplier strategy are selected simultaneously to compensate for shortages during moderate supply disruptions. The supply chain is completely dependent on external forces, and only the back-up supplier strategy is selected to maintain supply chain stability when the supply is severely disrupted.

5. Discussion

In order to further understand the influence of this mathematical model on the selection of the resilience strategy, the relevant parameters are analyzed, including the redundant inventory coefficient and the back-up supplier unit supply price coefficient.

5.1. The Impact of the Amount of Redundant Inventory

According to the actual situation, the number of prefabricated components from the first to the second floor of the prefabricated building project is kept as the inventory. It is rare for component supplier inventory to exceed 15%. Therefore, we stipulate that the other parameters remain unchanged, and the redundant inventory coefficient is set to 5%, 10% or 15% of the maximum supply capacity of each supplier.

To identify the solution that minimizes the total supply chain cost while maximizing resilience, we will focus on analyzing the optimized solutions that match the average value of the two objective functions. As shown in Figure 6, Figure 7 and Figure 8, the objective function values of Pareto solution sets with different redundancy coefficients are represented by line graphs, and the average values of the two objective functions are set as reference lines. The difference between the values of the two objective functions in each optimal solution and the average can be expressed intuitively.

When the redundancy coefficient is set at 5%, the second optimization scheme is considered to be the strategy closest to the average of the two objective functions. It is in scenarios 1, 2, 3 and 4 where back-up supplier strategies are selected to deal with supply disruptions. The redundancy coefficient of 5% means that the total redundant inventory of the three suppliers cannot meet the amount of stock shortage caused by any disruption scenario. Although the cost of redundant inventory management is low, only the redundant inventory strategy is selected, which does not satisfy the demand of the construction company. If the redundant inventory strategy and back-up supplier strategy are selected at the same time, the total cost of the supply chain will increase significantly. Therefore, only the back-up supplier strategy is selected to better balance the two objective functions. The purpose of enhancing supply chain resilience is achieved.

When the redundancy coefficient is 10%, the strategy that is closest to the average of the two objective functions is the 25th scheme. In scenarios 1 and 2, the redundant inventory strategy is selected; in scenario 3, the back-up supplier strategy is selected; and in scenario 4, both the redundant inventory and back-up supplier strategy are selected. When the redundancy factor is 15%, the strategy that is closest to the average of the two objective functions is the 11th scheme. In scenarios 1 and 2, redundant strategy is selected; in scenario 3, the back-up supplier strategy is selected; and in scenario 4, both the redundant inventory and back-up supplier strategy are selected.

The redundancy factor is 10–15%, which means that the number of redundant stocks can satisfy the number of stock shortages in some disruption scenarios. It is necessary to select a resilience strategy according to the degree of supply disruption in the scenario. In the case of mild disruption, the redundant inventory strategy is generally selected. Supply chains prioritize the use of their own capacity to absorb risk. When there is a moderate disruption, the supply chain will rely on external forces, and two strategies are chosen to maintain the stability of the supply chain. In the case of a severe disruption, it is a better decision to only select the back-up supplier strategy.

As shown in Figure 6, Figure 7 and Figure 8, when the redundancy inventory coefficient is set at 10%, there are significant fluctuations in the total supply chain cost and resilience. When the redundant inventory coefficient is 10%, the management cost of the redundant inventory is at a reasonable level. The difference between the start-up cost of the redundant inventory strategy and the start-up cost of the back-up supplier strategy is smaller. However, there is a large gap between the redundant unit supply cost and the unit supply cost of back-up suppliers. There is a significant difference in the total cost and resilience of the supply chain if different resilience strategies are selected. Similarly, when both strategies are selected, it is obvious that the total cost of the supply chain will change if the quantity selected in each strategy is different. It will further affect supply chain resilience due to the difference in supply quantity. Therefore, in order to enhance supply chain resilience, it is necessary for the general contractor to consider the various cost and resilience strategies, and the precise control of the supply quantity under different strategies also needs to be focused on. In addition, when the redundant inventory coefficient is set to 15%, there is a minimum value for the total cost of the supply chain compared to 5% and 10%. This is because the amount of redundant inventory is large enough that the general contractor can choose only the redundant inventory strategy to deal with the adverse consequences of supply disruptions, thus reducing the cost of choosing the back-up supplier strategy. However, from the perspective of the comprehensive resilience objective function, the redundant inventory level of 15% is not optimal. Therefore, the general contractor needs to optimize inventory levels based on the level of disruption and the desired level of supply chain resilience.

In summary, it can be concluded that when the quantity of redundant inventory can meet the amount of stock shortage resulting from the disruption, internal capacity within the supply chain will be prioritized, with a redundant stock strategy being selected as an emergency response. If the amount of redundant inventory matches the shortage, both the redundant inventory and back-up supplier strategy are simultaneously employed to ensure supply chain stability. If the quantity of redundant inventory is insufficient to cover the shortage, the construction general contractor will select a back-up supplier strategy to restore supply chain resilience. Although redundant inventory can improve the resilience of the supply chain, it also increases the cost of inventory. Therefore, the construction general contractor needs to balance the cost of redundant inventory with the loss of out-of-stock inventory caused by disruption and find the optimal inventory level to maximize the supply chain benefit.

5.2. Impact of Unit Supply Prices of Back-Up Suppliers

According to market research, the unit supply price coefficients of the back-up supplier is set to be 1.10 times, 1.20 times and 1.30 times the average unit supply price of suppliers in the supply chain, and other parameters remain unchanged. Figure 9, Figure 10 and Figure 11 show line graphs of the two objective function values in the Pareto solution set under different back-up supplier unit supply price coefficients.

When the back-up supplier unit supply price coefficient is set to 1.1, the optimal scheme that is closest to the average of the two objective functions is the 21st scheme. In scenarios 1 and 2, the redundant inventory strategy is selected; in scenario 3, the back-up supplier strategy is selected; and in scenario 4, both the redundant inventory and back-up supplier strategy are selected. When the back-up supplier unit supply price coefficient is set to 1.2, the optimal scheme that is closest to the average of the two objective functions is the 96th scheme. It is in scenarios 1, 2 and 4 that the redundant inventory strategy is selected, and in scenario 3, the back-up supplier strategy is selected. When the back-up supplier unit supply price coefficient is set to 1.3, the optimal scheme that is closest to the average of the two objective functions is the 103rd scheme. It is in scenarios 1, 2 and 4 that the redundant inventory strategy is selected, and in scenario 3, the back-up supplier strategy is selected.

The results indicate that in scenario 1 and scenario 2, the redundant inventory strategy is always selected by the general contractor no matter how much the price coefficient increases. Since the unit supply price of back-up suppliers increases, the cost of choosing a back-up supplier strategy to absorb the loss of supply disruptions will increase. However, in scenario 3, a back-up supplier strategy is selected to maintain supply chain stability. Because the average disruption degree of scenario 3 is the most severe, only the redundant inventory strategy selected cannot make both objective functions optimal. In addition, the cost of selecting two strategies to maintain the stability of the supply chain is too expensive. In scenario 4, the back-up supplier strategy is first selected. With the increase in the unit supply price coefficient of the back-up suppliers, the construction general contractor tends to select the redundant inventory strategy.

We can conclude that regardless of the change in the unit supply price coefficient of the back-up supplier, in the case of a minor disruption, the general contractor always chooses the redundant inventory strategy. When the degree of disruption is serious, a back-up supplier strategy is always selected to ensure the stability of the supply chain. However, when the supply chain is moderately interrupted, with the increase in the unit supply price coefficient of the back-up supplier, the general contractor will usually switch from the back-up supplier strategy to the redundant inventory strategy to restore the supply chain’s resilience level. Therefore, when managing supply chain disruptions, the general contractor possesses the flexibility to adjust the strategy according to the extent of the disruption and the unit supply price coefficient of the back-up supplier. This flexibility ensures that the supply chain can quickly adapt and restore stability in the face of different disruptions. Furthermore, the unit supply price coefficient of the back-up supplier can also facilitate collaboration between upstream and downstream enterprises in the supply chain. Through sharing price information and resources, companies collectively address supply chain risks and enhance overall resilience.

6. Conclusions

This paper considers two resilient strategies to promote the sustainable development of prefabricated building supply chains: redundant inventory and back-up suppliers. Aiming at minimizing the total cost and maximizing supply chain resilience, a supply chain resilience strategy selection model for different disruption scenarios was established. The NSGA-Π was proposed, and the combination of the PC and PM parameters in the algorithm was optimized and analyzed according to the project case. Furthermore, the impact of the redundant inventory coefficient and the unit supply price coefficient of the back-up supplier on the model was analyzed. The results indicate that the supply chain with resilience construction has a superior capability to cope with disruption. The flexible selection of resilience strategies according to varying degrees of supply disruption and the unit supply price coefficient of the back-up supplier, as well as the optimization of redundant inventory levels, is conducive to reducing the cost of resilience recovery after supply chain disruption, which further verifies the effectiveness of the model and the impact of uncertain parameters in the model on the results. Therefore, in the process of building a resilient supply chain, the general contractor should focus on the effective configuration and utilization of resources to avoid unnecessary waste so as to maximize the economic benefits while ensuring the stability of the supply chain.

It is necessary to continue to study how to enhance the resilience of the prefabricated building supply chain to achieve sustainable development when the supply chain is subject to different disruption scenarios. Just-in-time assembly, the transportation route and other constraints are not considered in this paper. In the future, more constraints should be considered when studying ways to enhance supply chain resilience. This will contribute to enhancing the management level of the supply chain of the construction general contractor, thus promoting the development of the construction industry to meet the requirements of sustainable development.