Data-driven Simulation-based Framework for Efficient Scheduling of Industrial Construction Projects
Uložené v:
| Názov: | Data-driven Simulation-based Framework for Efficient Scheduling of Industrial Construction Projects |
|---|---|
| Autori: | Taghaddos, Maedeh |
| Informácie o vydavateľovi: | University of Alberta Library, 2023. |
| Rok vydania: | 2023 |
| Predmety: | Optimization, Workface Planning, Industrial Construction, Scheduling, Flexible Resource Allocation, Simulation, 12. Responsible consumption, Space Management |
| Popis: | Efficient scheduling and resource allocation for large-scale industrial projects is challenging due to the size, labour intensity and complexities associated with such projects—particularly for fast-track contracts characterized by the unavailability of detailed information in the early planning phase. Thus, initial industrial construction schedules are often developed based on the subjective experiences of practitioners that are often overly ambitious, resulting in compressed timelines that require increased costs to achieve later. Moreover, space is a vital asset in congested industrial sites, which involve various constraints. However, many of such limitations are not considered in early planning, facing a lack of detailed information. In other words, traditional planning often disregards the inter-dependencies between schedule and space problems. Therefore, addressing site congestion during construction requires subjective ‘on-the-ground’ adjustments. This study first, introduces a comprehensive, hybrid planning framework capable of developing an efficient, high-level schedule accounting for uncertainties in planning and execution. The framework employs simulation modelling and optimization tools to determine the best option of durations/execution modes in multi-mode activity networks. The proposed multi-mode resource allocation framework uses historical information and expert opinion to create a near-optimized schedule while considering various uncertainties automatically. The proposed approach is tested and demonstrated with an illustrative example and an actual case study of industrial construction in Alberta, Canada. In the second phase of this research study, a multi-mode resource allocation approach is further improved and expanded to develop a variable and data-driven resource allocation framework by incorporating historical S-curve boundaries. The introduced approach offers variable, congestion-based resource allocation for early project planning through an automated, simulation-based scheduling framework to enhance high-level schedule quality. The developed optimization engine determines the variable resource allocation and is expected to profoundly affect some scheduling factors, such as float and resource fluctuation. The novel data-driven workface planning framework for industrial projects can respect various scheduling uncertainties and constraints (e.g., space). The developed framework is illustrated with a manual solution through a simple illustrative example. This system is also tested in an actual case study, demonstrating its functionality and improvements when applied to an industrial project. Phase 3 of this research proposes an evaluation procedure, which has been used to evaluate the proposed variable resource allocation framework. This procedure benefits the decision makers who need quantifiable schedule evaluation for various what-if scenarios assessments. First, evaluation metrics are introduced for schedule assessment. Next, synthetic data are utilized for project generation. This step is constructive for comprehensively evaluating the lack of sufficient actual data. The proposed variable framework has been evaluated by analyzing the results collected from generated synthetic projects along with the actual project according to the proposed evaluation metrics. The results show the improvement in all the proposed metrics, demonstrating the benefit of using the proposed variable resource allocation framework developed in Phase 2. |
| Druh dokumentu: | Thesis |
| Jazyk: | English |
| DOI: | 10.7939/r3-y4wm-1h77 |
| Prístupové číslo: | edsair.doi...........a7152e902b1ee208965373cb89dbb5d6 |
| Databáza: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | Efficient scheduling and resource allocation for large-scale industrial projects is challenging due to the size, labour intensity and complexities associated with such projects—particularly for fast-track contracts characterized by the unavailability of detailed information in the early planning phase. Thus, initial industrial construction schedules are often developed based on the subjective experiences of practitioners that are often overly ambitious, resulting in compressed timelines that require increased costs to achieve later. Moreover, space is a vital asset in congested industrial sites, which involve various constraints. However, many of such limitations are not considered in early planning, facing a lack of detailed information. In other words, traditional planning often disregards the inter-dependencies between schedule and space problems. Therefore, addressing site congestion during construction requires subjective ‘on-the-ground’ adjustments. This study first, introduces a comprehensive, hybrid planning framework capable of developing an efficient, high-level schedule accounting for uncertainties in planning and execution. The framework employs simulation modelling and optimization tools to determine the best option of durations/execution modes in multi-mode activity networks. The proposed multi-mode resource allocation framework uses historical information and expert opinion to create a near-optimized schedule while considering various uncertainties automatically. The proposed approach is tested and demonstrated with an illustrative example and an actual case study of industrial construction in Alberta, Canada. In the second phase of this research study, a multi-mode resource allocation approach is further improved and expanded to develop a variable and data-driven resource allocation framework by incorporating historical S-curve boundaries. The introduced approach offers variable, congestion-based resource allocation for early project planning through an automated, simulation-based scheduling framework to enhance high-level schedule quality. The developed optimization engine determines the variable resource allocation and is expected to profoundly affect some scheduling factors, such as float and resource fluctuation. The novel data-driven workface planning framework for industrial projects can respect various scheduling uncertainties and constraints (e.g., space). The developed framework is illustrated with a manual solution through a simple illustrative example. This system is also tested in an actual case study, demonstrating its functionality and improvements when applied to an industrial project. Phase 3 of this research proposes an evaluation procedure, which has been used to evaluate the proposed variable resource allocation framework. This procedure benefits the decision makers who need quantifiable schedule evaluation for various what-if scenarios assessments. First, evaluation metrics are introduced for schedule assessment. Next, synthetic data are utilized for project generation. This step is constructive for comprehensively evaluating the lack of sufficient actual data. The proposed variable framework has been evaluated by analyzing the results collected from generated synthetic projects along with the actual project according to the proposed evaluation metrics. The results show the improvement in all the proposed metrics, demonstrating the benefit of using the proposed variable resource allocation framework developed in Phase 2. |
|---|---|
| DOI: | 10.7939/r3-y4wm-1h77 |