Content area
In recent years, China’s construction industry has developed rapidly, but it still faces problems such as complex processes, long cycles, and unpredictable environments in engineering projects. To address these issues, a building information modeling development and design based on visual management was proposed to improve the efficiency of collecting and analyzing engineering data and information. At the same time, the particle swarm multi-objective optimization algorithm was adopted to comprehensively analyze the influencing factors during the operation of the module. The results indicated that the response time of the building information model in information processing did not exceed 20% of the total time. By using this method, the project cost prediction error reduced to 30–80 yuan, which demonstrated the accuracy of building information modeling technology. Compared with the efficiency value of the algorithm in the first 10s, the traditional single objective optimization algorithm was 0.28, while the proposed algorithm was 0.40. This significant improvement indicated that the development of building information models could effectively improve the efficiency of information flow. The particle swarm multi-objective optimization algorithm performed well in reducing project cost prediction errors. The results of this study have promoted the information process of the construction industry and provided strong support for achieving efficient and accurate engineering management.
Introduction
The market size of China’s construction industry is constantly expanding, and technological innovation and green development have become trends. With the development of big data, Internet of Things, artificial intelligence and other technologies, the construction industry will be transformed in an intelligent way. Intelligent buildings, intelligent construction sites, etc. will become new trends in the future construction industry [1]. Building information modeling (BIM) has been widely used as an efficient digital tool in building design, construction, and operation management. For example, Datta et al. evaluated the effectiveness of BIM applications by developing a 3D model of a three story residential building and integrating all semantic data to analyze the role of BIM in the planning and construction phases. The results indicate that BIM effectively improves the overall performance of the project, enhances communication efficiency, and reduces errors [2]. However, existing BIM models still face many challenges and limitations when dealing with multi-objective optimization problems. Traditional optimization methods often struggle to simultaneously address multiple conflicting objectives such as cost, schedule, and quality. It affects the practicality and reliability of the optimization results [3]. In this context, there is a need for better technical management and research in engineering projects. This can better calculate and analyze the various environmental conditions during the construction process, as well as the progress, cost, and various data information of the project [4]. In recent years, particle swarm optimization (PSO) algorithm, as an optimization algorithm for swarm intelligence, has the advantages of easy implementation, high accuracy, and fast convergence. Abualigah and other scholars compared the applications of PSO in energy systems, engineering design, and robotics, and the results showed that PSO has multifunctionality and effectiveness [5]. Although PSO has shown promise in structural optimization, previous research has unresolved issues such as single objective bias, insufficient BIM integration, and inability to handle discrete building variables. Therefore, this study innovatively combines visual management technology with multi-objective optimization algorithms and proposes a new management method. Firstly, a BIM based architectural information management model was proposed, and the development process of BIM models and their application in engineering projects were designed. Secondly, a bidirectional parameter mapping mechanism has been developed to automatically convert BIM model parameters into PSMOO parameters. PSMOO adopts Von Neumann structure and designs Pareto frontier accelerator. The implementation of visual management technology through BIM facilitates the analysis and processing of diverse engineering construction data, thereby enhancing the efficacy of information exchange and elevating the technical management level of engineering projects. Meanwhile, the traditional PSO has been improved by transforming it into the particle swarm multi-objective optimization (PSMOO) algorithm. This can better handle the search problem of discrete variables in engineering and improve the effectiveness of optimization algorithms. Ultimately, PSMOO will be combined with BIM models to find the optimal solution considering multiple design and construction objectives. The research aims to better support architectural design and construction decisions, provide an efficient and reliable method for project management, and improve the overall effectiveness of project management.
Four sections make up the research. The first section discusses the plans for domestic and international research in PSO and visualization. The second section is about the development of BIM module and the structure of PSMOO algorithm as well as the algorithm according to the process. The third section tests the models and algorithms improved and optimized by the research. The fourth section is the summary discussion of the test results and the analysis of the shortcomings of the study.
Related work
With the rapid development of construction information technology, it is important to build a reasonable information exchange platform to study the environment as well as the technology of the project. BIM has advantages in improving design quality, optimizing construction plans, and reducing project costs in engineering project management. Among them, Waqar et al. combined exploratory factor analysis and partial least squares structural equation modeling to identify the key success factors of BIM application in engineering projects. The results showed that BIM could improve project efficiency, cost savings, and multi-party collaboration [6]. To identify the main parameters affecting the rational management of soil, Borkowski et al. investigated using a novel BIM technology model. The results showed that the model could be effective in obtaining data on soil types and/or agricultural use complexes [7]. Al Roumi et al. developed a standardized 4D BIM template for residential buildings using process pattern recognition methods to address the issue of BIM application gaps in Kuwait. The results showed that this method had good practicality [7]. Nguyen et al. conducted exploratory factor analysis and partial least squares structural equation modeling to identify the six key factors promoting BIM adoption in construction companies and their relationships. The results showed that BIM effectively achieved digital transformation [8].
PSO is widely used in the field of construction project management to achieve optimization goals, including full-cycle management processes such as design, cost, budget, and schedule. Liu et al. developed a Python program based on an improved PSO algorithm to solve the optimization design problem of slope support. The key parameters of anchor rod lattice beams in slope engineering were optimized. The results showed an average improvement of 30.5% in slope stability [9]. Almahameed et al. integrated PSO and voting regression methods to achieve precise cost control and resource allocation in engineering projects. This integration was undertaken to optimize and predict construction costs. The results showed that this method was effective [10]. Arzanlou et al. constructed an allocation optimization model for project portfolio budget optimization problems through mathematical modeling and PSO algorithm, and solved the optimal allocation scheme under operational and budget constraints. The results showed that the method was efficient [11]. To improve the spatio-temporal prediction accuracy of engineering projects, Bakhshi et al. combined the earned value management method with the least squares support vector machine-PSO hybrid algorithm to verify the data of dams under construction. The results showed that the algorithm had good performance, low error variance, and strong generalization ability [12].
In summary, although many researchers have extensively studied BIM and PSO, demonstrating their advantages in their respective fields, there are few design concepts that combine the two. Their combined impact on practical engineering project management has not been fully demonstrated. Therefore, the study develops the BIM module to address the problem of difficult to collect and analyze data and information in engineering construction. The research also improves PSO to optimize the solution of discrete variables that occur in the process of project management. It is expected that the algorithms proposed by the research can have better applications in engineering and technology management.
Engineering project management based on BIM model and PSO optimization
The study is divided into two subsections to investigate the visualization simulation module and PSO for engineering project management techniques. First, the study designs the model and development logic for the integration of BIM and engineering projects. Then the PSMOO model structure is developed.
Engineering projects based on BIM model development
Building informationization model based on BIM provides a platform for information exchange and sharing among different stages of the whole life cycle of engineering construction and different participating subjects [13– 14]. Information in the process of engineering construction is redundant, not easy to deal with, and very likely to affect the progress of the project. The study revised the informationization model in an attempt to hasten the pace of information exchange and raise the engineering construction industry’s degree of productivity. The developed BIM model construction industry informationization approach is shown in Fig. 1.
[See PDF for image]
Fig. 1
BIM based construction informationization
In Fig. 1, BIM is responsible for visual simulation, temperature analysis, environmental analysis, sunshine analysis, etc. In BIM, dynamic parameters can be obtained through sensors, and effective real-time data can be collected. BIM is also needed to regulate schedule control, cost budgeting, quality and safety management, and facility maintenance during the course of the project. Such regulation is used to support the technology and methodology of decision-making, design, construction, and operation of BIM building full life cycle projects. The BIM model development logic diagram is shown in Fig. 2.
[See PDF for image]
Fig. 2
BIM model development logic diagram
In Fig. 2, the BIM model includes multiple influencing factors such as construction, installation, structure, unit, cost, and schedule. By analyzing multiple influencing factors through BIM models, the model can be controlled and information feedback can be achieved through the project application. The first application of the model is used to test the reliability of the model and to evaluate and demonstrate the performance and feasibility of the model. The second development of the model is mainly to improve the problems occurred in the first application, and then add optimization and design ideas. Controlling and analyzing the disruptive elements impacting the model application project is the third evolution of the model. After the development is completed, it can be applied to the collection of building information. The impact factor calculation by Eq. (1) [15].
1
.In Eq. (1), denotes the impact factor. and denote two similar impact factor events, respectively. Impact factor under combined influence of and , the combined impact factor can be expressed as Eq. (2) [16].
2
.In Eq. (2), denotes the representation of the impact factor under the joint influence of and . If any number of events are considered to occur simultaneously, the multi-factor effect can be expressed as Eq. (3) [17].
3
.In Eq. (3), , , , , and denote the coefficients of the influencing factors. denotes the set of influencing factors. If only the effect of any two factors is considered, the two-factor impact can be expressed in Eq. (4).
4
Similarly, when considers three influencing factors, the three factor impact can be expressed as Eq. (5).
5
.Once the BIM has been developed, particular tasks should be performed to confirm the model’s efficacy in the engineering implementation project. Figure 3 displays the model’s customized progress management flow chart.
[See PDF for image]
Fig. 3
BIM progress management process
In Fig. 3, the process of BIM schedule management is based on the principle of the last planner system of lean construction management theory and technology. Following the hierarchy of project master schedule, secondary schedule, weekly schedule, and daily work, the entire process pulls from the back of the process to the front of the process. Plans are summarized from the bottom up to ensure as little buffer stock as possible. The master schedule is completed first, after which the secondary schedule is subsequently prepared and tasks are refined, and work is coordinated to submit the work plan. If problems are encountered at the start of a task, the task is stopped and returned to the start of the plan, where the plan is discussed and changed. According to the schedule management plan, the target schedule can be expressed as Eq. (6) [18].
6
.In Eq. (6), denotes the target progress. denotes the predicted progress. If and are mutually independent events, their joint probability function can be expressed in Eq. (7).
7
.In Eq. (7), denotes the probability function of the influence factor . denotes the probability function of the influence factor . and denote the normal distribution graphs of two different events. denotes the expected value. The test data comes mainly from historical data and related databases of actual construction projects, including architectural design drawings, construction logs, material lists, maintenance records of equipment and facilities, etc. The collected data is cleaned and preprocessed to remove missing and outlier values. The normalization of continuous variable data uses the Min-Max normalization, which divides the preprocessed dataset into training and test sets in an 8:2 ratio.
Model structure based on PSMOO
The processed data is integrated into the BIM platform and optimized using the PSMOO algorithm. The development of BIM provides a lot of help in the whole process of project carrying out, but there is a drawback of low efficiency in information exchange. PSO uses information sharing and cooperation between individuals within a group to seek the optimal solution, which is mainly optimized for a single objective problem [19]. Engineering projects often involve multiple conflicting objectives such as schedule, cost, and resource optimization, requiring an optimization algorithm that can handle multiple objectives simultaneously. Therefore, this study introduced PSMOO. PSMOO can effectively handle trade-offs among multiple objectives by maintaining a set of non-dominated solutions (Pareto front) in the population, thereby finding the optimal solution that satisfies different objective requirements. PSMOO combines the advantages of both global and local models. In Fig. 4, the two models are displayed.
[See PDF for image]
Fig. 4
Local and global model model
In Fig. 4, the global model can find the global optimal solution faster by allowing all particles to share information and converge quickly. However, it is also more prone to getting stuck in local optimal (LO) solutions. When engineering projects have high time requirements, global models have advantages. On the contrary, local models can better avoid getting stuck in local optima by limiting the range of information transmission, but their convergence speed is slower [20]. When engineering projects require high accuracy of results, local models are more suitable. By combining the advantages of both, appropriate models can be selected in different engineering projects to improve the efficiency and effectiveness of optimization [21]. The flowchart of the PSMOO algorithm is shown in Fig. 5.
[See PDF for image]
Fig. 5
PSMOO algorithm process
In Fig. 5, the process begins with setting the relevant parameters of the algorithm. The parameters such as progress, cost, and resources are extracted from BIM models created using Revit or Autodesk Navisworks software. Input real-time data, then the population initialization is performed to initialize the position and velocity vectors of each particle itself. Then the adaptation values (AVs) of the particles are calculated. Following the update of the particle positions and velocities, the AV of each particle is computed and compared to the AV that corresponds to its previous ideal position. The current location is updated using if the AV is high at that moment [22]. The AV of every particle is contrasted with the AV that corresponds to the global ideal position. The current particle position is used to update the global optimal location if the current AV is high. Finally, the search results are tested to see if they satisfy the termination conditions. The search should be stopped and the current AV should be output as the global optimal solution if it does not satisfy. If it does, then return to continue the iteration. Among them, BIM parameters such as component quantity and duration are mapped to the PSO particle dimension. Equation (8) provides the expression for updating particle velocity after iterations.
8
.In Eq. (8), denotes the flight speed of the particle in space. denotes the inertia weight factor. It is used to measure the ability of particles to maintain their current velocity in the search space. adopts a linear decreasing strategy, with an initial setting of 0.9 to enhance global exploration and a later setting of 0.4 to enhance local development [23]. denotes the number of iterations. and denote the particle flying to the optimal position and the overall optimal position step, respectively, which are called learning factors. Both allow the particles to find a balance between global exploration and local exploitation. and denote random numbers. denotes the particle’s best position. denotes the global best position. Equation (9) provides the expression for the position of the th particle in the population after iterations.
9
In Eq. (9), denotes the population and determines the number of particles. denotes the position that the particle seeks on its own during the movement. The update of the position of the particle at the next update can be expressed in Eq. (10) [24].
10
10)In Eq. (10), denotes the number of iterations. denotes the particle coefficient. The position of the next iteration is denoted as Eq. (11).
11
The optimized objective function (OF) can be expressed as Eq. (12).
12
.In Eq. (12), denotes the OF of the initial position of the particle. denotes the OF of the particle motion position. Equation (13) represents the extreme situation.
13
.The attribute case of can be expressed in Eq. (14).
14
BIM progress, cost, and other OFs are converted into PSO fitness functions. Equation (15) can be used to express the OF and satisfy this requirement. OF stands for BIM progress and cost optimization [25].
15
.Update BIM model parameters after each PSO iteration and achieve bidirectional linkage through Dynamo scripts. The PSMOO flowchart ensures that the particles are able to find the optimal position with the optimal speed exactly as they move through the population. The topology of PSMOO can also have an impact on the performance of the algorithm to a great extent. The multiple topologies of the PSMOO are shown in Fig. 6.
[See PDF for image]
Fig. 6
Topology of PSMOO algorithm
In Fig. 6, several basic topologies of the algorithm are ring-type, random ring-type, wheel profile, random wheel profile, Von Normanyi structure, and random structure [26]. In the star structure, each individual is connected to all other members of the group. It is characterized by fast information transfer and fast convergence, but it is easy to fall into a LO. The ring-type structure is where all the particles are lined up in a ring-like structure. It is characterized by slow information transfer and slow convergence, but it is not easy to fall into a LO contrary to the characteristics of the star structure [27]. The Von Neumann structure is shown in Fig. 6 (e). This structure is a three-dimensional square structure that appears as a lattice. Each particle communicates with its connected particles above, below, left, and right. The distribution of particles inside the structure varies from one structure to another. By changing the structure, the centralized management of dispersed particles can be realized, and the most suitable structure can be matched by the information transfer of internal particles and the convergence speed. It performs best in the three-dimensional solution space due to its neighborhood connectivity theorem, which states that 4-neighborhood communication can balance exploration and development. At the same time, the application of structure in PSMOO is more suitable for high-dimensional mapping of BIM parameters, such as schedule, cost, and resource three-dimensional target space.
Performance test of BIM design development with PSMOO algorithm
To assess the effectiveness of the research, this chapter is split into two pieces. The first segment examines the simulation results and assesses how well the specified and created BIM model performs. The second section tests the performance of the PSMOO algorithm and analyzes it in comparison with other algorithms of the same type.
Performance testing and simulation effect analysis for BIM design development
The experimental hardware configuration is NVIDIA RTX A6000 GPU, BIM modeling is Autodesk Revit 2023 software, and parameterization is Dynamo. The PSMOO implementation software is Python 3.9. PSMOO sets the population size to 100, iteration times to 500, initial inertia weight to 0.9, later inertia weight to 0.4, learning factors to 1.5, and random numbers uniformly distributed in [0,1]. The data comes from five coastal infrastructure projects completed by Chinese transportation construction companies between 2019 and 2023, including original schedules, cost reports, construction logs, and other data. Table 1 displays comparison data of project progress simulation optimization findings. Data in the ‘Informationized Prediction’ column is derived from historical averages, while ‘BIM Prediction’ values are outputs from PSMOO-optimized simulations. Missing cost entries (/) indicate unavailable historical data due to project-specific contingencies.
Table 1. Simulation and optimization results of project schedule
Construction serial number | Project name | Time limit for construction (day) | Cost (Ten thousand yuan) | Construction serial number | Project name | Time limit for construction(day) | Cost (Ten thousand yuan) |
|---|---|---|---|---|---|---|---|
1 | Site preparation | 12 | 14 | 1 | Site preparation | 12 | 15 |
10 | Underwater reef blasting | 4 | 14 | 10 | Underwater reef blasting | 4 | 16 |
11 | Base groove dredging | 5 | 15 | 11 | Base groove dredging | 4 | 14 |
12 | Bedrock riprap | 5 | / | 12 | Bedrock riprap | 7 | 6 |
13 | Tamping and leveling | 8 | / | 13 | Tamping and leveling | 9 | 14 |
14 | Caisson mounting | 8 | / | 14 | Caisson mounting | 9 | 7 |
15 | The caisson is filled with sand | 7 | / | 15 | The caisson is filled with sand | 7 | 6 |
In Table 1, the duration prediction for construction site preparation has a duration of 12 days and a predicted project cost of 140, 000 yuan for the informationization prediction. The predicted duration of the BIM development model is consistent with the informationization prediction, and the gap on the cost prediction is 10, 000 yuan. In the process of underwater operation, the informatized prediction of duration is 4 days and the cost is 140, 000 yuan. BIM predicts a duration of 40, 000 days and a cost of 160, 000 yuan. In the dredging process, the informational forecast duration is 5 days and the cost is 150, 000 yuan. The BIM duration is 40, 000 yuan and the cost is 140, 000 yuan. In the process of rock throwing operation, the informational forecast duration is 5 days and the cost forecast is unknown. BIM predicts a duration of 7 days and a cost of 140, 000 yuan. Due to the complexity of the process, it leads to an unknown phenomenon in the cost prediction of the project by the general information technology. BIM is developed without unknowns. The forecasted duration for the leveling process is 9 days at a cost of 140, 000 yuan. The duration of installation is 9 days at a cost of 70, 000 yuan. The duration of sand filling is 7 days at a cost of 60, 000 yuan. To verify the effectiveness of BIM, its performance is compared with four information processing models, M1, M2, M3, and M4. A comparison of the response times of the different information processing models for the different phases of the project is shown in Fig. 7.
[See PDF for image]
Fig. 7
Comparison of response time of different information models in different stages of engineering
Figure 7a shows the changes in sensitivity of different models under a small amount of training time. When the project is in the initial stage, the model response times are all low, very sensitive to the information, and fast processing speed. The response time of the BIM model in the initial stage of the project is about 750 s for less number of training. The fastest response time among the other models is 900 s, and the BIM response time is 15% of the total response time. The response time of BIM at the middle stage is about 1200 s, and the response time is about 17% of the total response time. Figure 7b shows the variation of sensitivity of different models with increasing training time. The response time of BIM increases with the number of training sessions, but none of them exceeds 3000 s. In Fig. 7a, the response time of the BIM response time in the late stage is about 1300 s, while the fastest response time of the other models is 1550 s. The response time of this model in the project closing stage is still the fastest, while the fastest response time of the other models is 2100 s at the maximum. The response time of the BIM model in processing the information does not account for more than 20% of the total time in the whole process of the proceeding. The results before and after cost estimation using the BIM model in the project works are shown in Fig. 8.
[See PDF for image]
Fig. 8
Results before and after cost estimation using BIM model
In Fig. 8a, the error between the project’s predicted and actual cost values is larger before cost estimation using BIM. The maximum value of the error occurs in project 22. Before the cost estimation without BIM, the predicted value of the project is 2400 and the actual value is only 2200, while the cost error after the application of BIM is reduced to within 100. The cost error range of the project before the application of BIM is between 100 and 200 yuan. The error is significantly reduced after cost estimation using BIM, as displayed in 8(b). At this point, the predicted value of cost in project 25 is 2420 and the actual value is 2470. The error range of the project cost after the implementation of BIM is 30–80 yuan. The error between the actual and predicted values of the project cost is further reduced. Through paired sample t-test, there was a significant difference in cost estimation error before and after BIM application (t = 6.32, p < 0.001). Cohen’s d = 1.87, and a value greater than 0.8 is considered a large effect, indicating that BIM intervention has practical application value.
Performance testing and simulation effect analysis of PSMOO algorithm
To verify the performance of the PSMOO in this research, a comparison experiment is designed. The PSMOO is compared with the conventional algorithm as well as the same type of algorithm in the same experimental environment. Table 2 displays the amount of time needed to process the various engineering materials used in the experiment as well as their quality criteria.
Table 2. Relevant requirements for engineering materials in simulation experiments
Construction process | Time limit for construction (day) | Quality index |
|---|---|---|
Structural concrete | 7 | Strength, compactness, crack resistance, durability, slump |
Steel structure construction | 5 | Welding quality, installation accuracy, length deviation, section deviation, levelness, surface quality |
Prefabricated column foundation | 14 | Bearing capacity, verticality, levelness, void rate, concrete strength, welding quality |
Fire protection engineering | 14 | System reliability, fire extinguishing, safety, anti-interference, construction quality |
Roof waterproofing | 5 | Waterproof thickness |
Table 2 shows the requirements of the specific engineering materials needed for the experiment. Structural concrete requires high strength, high compactness, and low collapsibility while ensuring water secretion, durability, and heavy water-to-cement ratio. The steel structure requires welding quality, installation precision, length accuracy, section deviation, level verticality, and surface quality. Grouted piles require bearing capacity, verticality, end face shape as well as horizontality, void rate, concrete strength, and welding quality. Fire engineering to ensure system reliability, fire extinguishing effect, anti-interference, ease of operation and construction quality. The waterproofing aspect of the level ensures the density, strength, elongation, durability, aging resistance, and bonding fastness of the waterproofing layer. The algorithm’s performance is evaluated under these engineering material settings. The results of the comparison between the PSMOO and other algorithms in the same environment are shown in Fig. 9.
[See PDF for image]
Fig. 9
The comparison results between the PSMOO and other algorithms in the same environment
In Fig. 9, the determined value of the traditional algorithm (TA) increases when the evaluation value is 0.04–0.20. After the evaluation value is greater than 0.20, the determined value of TA also keeps decreasing, and its effect is very poor. When the evaluation value is 0.22–0.37, the determined value of Algorithm 1 keeps increasing. The deterministic value of Algorithm 1 also continues to decrease when the evaluated value is greater than 0.37. At this point, Algorithm 1 has less effect. When the evaluation value is 0.40–0.56, the determined value of Algorithm 2 increases. After the evaluation value is greater than 0.56, the determined value of Algorithm 2 also decreases. At this point the algorithm works better. When the evaluation value is 0.56–0.74, the determined value of PSMOO continues to increase. After the evaluation value is greater than 0.74, the certainty value of PSMOO continues to decrease, and its effect is ideal. Figure 10 displays the test of efficiency values for various methods in the same test setting.
[See PDF for image]
Fig. 10
The test situation of efficiency values of different algorithms in the same test environment
In Fig. 10, for the first 10s, the efficiency value of the TA is around 0.28, the efficiency value of Algorithm 1 is around 0.37, and the efficiency value of the PSMOO is around 0.4. When compared to the previous algorithms, the enhanced algorithm’s efficiency value has increased with time. The efficiency values of the classic information algorithm, Algorithm 1, and the PSMOO are all approximately 0.38, 0.55, and 0.38 at the fifth 10s, respectively. When the efficiency value of the TA is unchanged, the efficiency value of Algorithm 1 is unchanged at 0.38 and the efficiency value of the PSMOO is around 0.4. At the last 10s, the efficiency value of the TA is around 0.38, the efficiency value of Algorithm 1 is 0.39, and the efficiency value of the PSMOO is around 0.49.
Discussion
To improve the overall effectiveness of project management, this study proposed BIM technology for visual management and used PSMOO to determine the optimal solution for multiple design and construction goals. The results indicated that the information processing time of the BIM model in this study was less than 20%, which was better than the industry benchmarks reported in reference [7] and others. It averaged 35%, but was slightly lower than the ideal value of 10–15% proposed in reference [22]. This might be due to the fact that the model used in this study did not integrate real-time IoT data. The proposed PSMOO evaluation value had an inflection point of 0.74, which was similar to the slope optimization threshold observed in reference [9]. In contrast, the multi-objective conflict handling capability of PSMOO was superior to the budget al.location model in reference [11], which was attributed to the introduction of a sensitive information feedback system. In this study, the project cost error was reduced to 30–80 yuan, which was equivalent to a budget deviation of ± 2.5%. This could provide project managers with a more accurate cash flow warning window. Moreover, this technology played a good role, similar to the ± 3% threshold effect verified in the dam project in reference [12]. In addition, the method of combining BIM and PSMOO could compensate for the deficiency of “lack of dynamic linkage between BIM adoption factors” pointed out in the reference [8], and achieve multi-objective real-time optimization of design construction cost. However, in PSO modeling, the weights of each objective were assumed to be linearly adjustable without considering the nonlinear effects of construction emergencies. Furthermore, compared with reference [28], the current BIM module did not integrate real-time geological monitoring data. In the future, GIS dynamic flow data needed to be integrated.
In summary, this study proposes a novel multi-objective optimization tool for the construction industry, thereby promoting the further development and popularization of BIM technology in practical applications.
Conclusion
To enhance the efficacy of information management in engineering projects, this study proposed a decision support system that integrates BIM visualization and PSMOO. This system enabled the real-time optimization of design and construction cost in three dimensions. In the experiments on the response time of the model to the information, the response time of the BIM model at each stage of the project was 15%, 17%, 16%, and 17% of the total time, respectively. The response time of the BIM model to process the information as a percentage of the total time did not exceed 20% throughout the process. In the process project cost prediction experiment using the developed BIM model, the prediction error of project cost was controlled within a budget deviation of 30–80 yuan. The evaluation value of PSMOO gradually increased within the determined range of 0.56–0.74. After the evaluation value exceeded 0.74, the evaluation value of the algorithm continuously decreases. The tipping point of the PSMOO evaluation value was increased to 0.74, which was significantly better than the traditional single objective optimization. Comparing the efficiency values of the algorithms, in the first 10s, the efficiency value of TA was 0.28 and the efficiency value of PSMOO was 0.40. Regardless of the time period, the efficiency value of PSMOO was always at its maximum. The proposed method could achieve multi-objective optimization and improve the efficiency of engineering project management. In conclusion, the development of BIM modules and optimized multi-objective algorithms is helpful for engineering research and analysis of engineering projects. Although this study has made some progress in optimizing construction project management, in practical applications, external uncertainties such as weather disturbances and supply chain issues often have a significant impact on construction progress and costs. Future research will focus on considering these external uncertainties, and further improve the adaptability and robustness of the model by introducing real-time environmental data such as weather information and supply chain status, as well as fuzzy logic processing techniques. In addition, combining advanced technologies such as federated learning to solve multi project collaborative optimization problems can enhance the broad applicability of research.
Acknowledgements
Not Applicable.
Author contributions
Hua Tian wrote the manuscript text.
Funding
No funding.
Data availability
The datasets generated during and/or analysed during the current study are not publicly available due [REASON(S) WHY DATA ARE NOT PUBLIC] but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1. Mohamed, AG; Marzouk, M. A BIM-based semantic approach for fund-allocation of Building components. Architectural Eng Des Manage; 2024; 20,
2. Datta, SD; Sobuz, MHR; Mim, JNAD. Investigation on the effectiveness of using Building information modeling (BIM) tools in project management: a case study. Revista De La Construccion; 2023; 22,
3. Dong, T; Li, H; Zhang, Z. The using effect of fuzzy analytic hierarchy process in project engineering risk management. Neural Comput Appl; 2025; 37,
4. Vinod, PI; Mahesh, G. Developing a project performance assessment model for benchmarking the project success of Indian construction projects. Benchmarking: Int J; 2024; 31,
5. Abualigah, L. Particle swarm optimization: advances, applications, and experimental insights. CMC-COMPUTERS Mater CONTINUA; 2025; 82,
6. Waqar, A; Othman, I; Almujibah, H et al. Implementing Building information modeling (BIM) for the success of geotechnical offshore construction projects: Malaysian construction industry. Qual Quant; 2024; 58,
7. Al-Roumi, H; Al-Sabah, R. Exploring the rate of adoption and implementation depth of Building information modeling (BIM): A case of Kuwait. J Eng Res; 2024; 12,
8. Nguyen, TT; N, Do, ST; Nguyen, VT et al. Interrelationships among enabling factors for BIM adoption in construction enterprises. Eng Constr Architectural Manage; 2024; 31,
9. Liu, Y; He, J; Liang, J et al. A new improved particle swarm algorithm for optimization of anchor lattice beam support structures. Nat Hazards; 2025; 121,
10. Almahameed, B; Bisharah, M. Applying machine learning and particle swarm optimization for predictive modeling and cost optimization in construction project management. Asian J Civil Eng; 2024; 25,
11. Arzanlou, A; Sardroud, JM. Enhancing project performance: particle swarm optimization for optimal budget allocation and maintenance scheduling. KSCE J Civ Eng; 2024; 28,
12. Bakhshi, R; Fard Moradi Nia, S; Jani, R et al. Applying project management knowledge and hybrid algorithm in predicting the time and cost of completing dam projects. Water Resour Eng; 2024; 17,
13. Jianhua, D; Yanping, W. Application of lévy flight particle swarm optimisation in MPPT of photovoltaic system. Int J Electron; 2024; 111,
14. 2Das Bed Prakash Kaushik, DS; Amitava, C. Bera Jitendra nathl. PSO-guided optimal estimator enabled regularized adaptive extended Kalman filter with unknown inputs for dynamic nonlinear indoor thermal state Estimation. J Build Perform Simul; 2024; 17,
15. Ye, Z; Antwi-Afari, MF; Tezel, A et al. Building information modeling (BIM) in project management: a bibliometric and science mapping review. Eng Constr Architectural Manage; 2025; 32,
16. Nguyen, TA; Nguyen, TA. Building information modeling (BIM) for construction project schedule management: a review. Eng Technol Appl Sci Res; 2024; 14,
17. Pande, PB; Dhengare, SW; Raut, JM et al. Optimizing sustainability and resilience of composite construction materials using life cycle assessment and advanced artificial intelligence techniques. Asian J Civil Eng; 2025; 26,
18. Tian, L; Wright, A; Painter, B et al. Factors influencing BIM use in green Building construction project management in the UK and China. Building Res Inform; 2023; 51,
19. Si, J; Wan, C; Hou, L et al. Self-organizing optimization of construction project management based on Building information modeling and digital technology. Iran J Sci Technol Trans Civil Eng; 2023; 47,
20. Himani, P. Multi-objective optimization of a sequentially inspected system prone to revealing and non-revealing failures via particle swarm optimization and a fuzzy method. J Industrial Prod Eng; 2024; 41,
21. Liang, R; Ma, H; Wang, P et al. The applications of Building information modeling in the life-cycle of green Buildings: a comprehensive review. Sci Technol Built Environ; 2024; 30,
22. Parsamehr, M; Perera, US; Dodanwala, TC et al. A review of construction management challenges and BIM-based solutions: perspectives from the schedule, cost, quality, and safety management. Asian J Civil Eng; 2023; 24,
23. Zijing, Y; Huan, L; Wenhong, W. Improved particle swarm optimization based on multi-strategy fusion for UAV path planning. Int J Intell Comput Cybernetics; 2024; 17,
24. Jiang, H; Jiang, A. An integrated system for tunnel construction safety control based on BIM–IoT–PSO. J Civil Struct Health Monit; 2024; 14,
25. Wu, Y. Design and analysis of energy efficient urban buildings based on BIM model. J Comput Methods Sci Eng; 2024; 24,
26. Mohammad Hasani Zade, B; Mansouri, N; Javidi, MM. Multi-objective task scheduling based on PSO-Ring and intuitionistic fuzzy set. Cluster Comput; 2024; 27,
27. Singh, R; Sangeeta, V; Manjunath, TC et al. MOPSO-driven optimization for sustainable retrofitting: balancing time, cost, and environmental impacts. Asian J Civil Eng; 2025; 26,
28. Borkowski SA, Bielska A, KwiatkowskaMalina J. Site-specific efficient management of soil resources using GIS and BIM technologies. Rep Geodesy Geoinformatics 2024,117(1):89–98.
© The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.