Efficient multiple project management is very important to the project-based industries. Current multi-project studies in shipbuilding focus on designing mathematical models and heuristic algorithms to achieve optimal resource usage. However, due to the rigid requirements on complete information, these models are inapt to support decisions in the early stages (such as the project biding stage) that have been acknowledged more and more critical in the ever fierce market. On the other hand, although pieces of information management software have been developed for project management and production, there are works left to be manually executed, such as production prediction of a new project and control of the temporary system access delegated to business partners, hindering further improvements in work efficiency. To bridge these gaps, this paper innovatively proposes a governance platform architecture based on the theory of Governance of Projects. The framework views organization management as important as project management and adopts a new access control method which helps reduce the manual labor. Moreover, a case-based reasoning algorithm that supports planning prediction with limited information is designed. Finally, a prototype system is developed and tested in a shipyard in China. It proves to be both effective and efficient.

Introduction

Shipbuilding industry, characterized by the ETO (Engineering-toOrder) production mode, is typically a project-based industry (Zhang et al., 2012). To compete in the ever fierce market, shipbuilding companies are demanded to be competent to conduct several projects at the same time (Ahola & Davies, 2012). Generally, a shipbuilding project is very large and complex (Han et al., 2017). The number of activities can reach up to ten thousand and cross-working is common in shipyards. Since key resources (such as docks, quays and so forth) are finite and have a direct impact on the final project makespan, it is a primary issue to achieve continuous production with minimum project delay in the multi-project context. Meanwhile, considering risks brought by the high budget and the long period of a shipbuilding project, managers always tend to get a more precise prediction and estimation of a new project to bid. Majority of current multi-project studies in shipbuilding focus on designing mathematical models and heuristic algorithms to solve the Resource Constrained Project Scheduling Problem (RCPSP) (Han, Yang, & Gong, 2010; Li, Hu, Lv, & Sun, 2013a, 2013b). Zhang et al. (2012) extends the RCPSP model with budget constraints and develops an optimization engine on top of Microsoft Project Server. However, rigid requirements on full information make these models inapt to support decisions in the early stages (such as the project biding stage). Boer (1998) develops a decision support system for shipyards to plan and schedule multi-projects. Although RCCP (Rough Cut Capacity Planning) function is provided in the order acceptance phase, the system still subjects to the input acquisition (a large number of process planning data) difficulties. Hans, Herroelen, Leus, and Wullink (2007) proposes a hierarchical project planning framework where flexible usage of planning methods is favored. Nonetheless, the research does not develop any practical tools. Some researchers indicate that planning is a knowledge-intensive work and introduce case-based reasoning (CBR) method in process planning (Cho, Lee, & Chung, 1996; Qu, Jiang, & Tao, 2013; Seo, Sheen, & Kim, 2007) as well as in work structure decomposition (Li, Mao, & Zhang, 2017) for shipbuilding projects. It is noticeable that a whole plan of a potential project cannot be generated by these studies, let alone a balanced result based on the current production status (Loaggui, Lu, and Xu (1987) develops a network-planning aid system for multi-shipbuilding production. The system can support rapid project duration estimation of a new project on condition that the process sequence is known. In other words, decision makers have to acquire the process sequence of a new project by other tools (or manually). To conclude, there is a lack of an automatic tool to predict production for decision makers in the biding stage of a shipbuilding project. On the other hand, more and more information management systems have been particularly developed for ship manufacturers to well manage their resource, project schedule, production cost and suppliers. Although such systems prove to be effective in their perspective domains, workers complain that sometimes they need to use several systems to complete one business workflow. It can be time-consuming and error-prone.