Norwegian case study: The role of operations research and behavioral operations management in handling project complexity.
By Hajnalka Vaagen
Design uncertainty is the major driver of planning complexity in building specialized vessels used in offshore oil and gas exploration.
“Built on trust” is the motto of the VARD Group AS, one of the most innovative global designers and builders of specialized vessels used in offshore oil and gas exploration. Despite being located in Norway – a high-cost country with suboptimal infrastructure and employment conditions – the company ranks among the best in terms of flexibility, delivery precision and short delivery times at competitive prices. In this environment, project work is the norm, and responsiveness to frequent technical changes is critical for performance.
“Trust” is undoubtedly an important competitive element of the Norwegian industrial culture. Given that, planning and scheduling large, complex projects normally requires advanced decision support tools, and operations research is expected to play a vital role in this task.
The peculiarities of shipbuilding for the oil and gas market segment, however, make planning very difficult, and the role of model-based decision aids is limited when it comes to handling the true complexity. In this environment, vessels may change substantially from contracting to delivery due to frequent and unsystematic client input, as well as by frequent regulatory interventions, all while working on the edge of known technology [3]. Since short delivery times are critical, almost every ship is put into engineering and production before all the technical uncertainty is resolved.
Figure 1: Comparison of the expected total costs of the reactive (deterministic) and proactive (stochastic) strategies.
The engineering of the vessel commonly starts when only the footprint of a strategic component is known, and large-scale strategic adaptations, far into the production process, may also occur. The planning complexity, therefore, arises from frequent design changes and advanced design and engineering taking place concurrently with production. Concurrency is challenging only if design is uncertain, which makes “design uncertainty” the major critical element of the planning problem; future decisions (on how to manage project task and resources) are conditioned on the future realization of the uncertain design. Despite this, design is often separated from project scheduling [3], mainly because ship design and engineering are still considered the domain of naval architects and engineers.
Before design and scheduling can be integrated, it’s first necessary to know how to handle uncertainty and dynamics (that is, the steady arrival of new information) in pure project scheduling for a given design. Finding a way to formulate this stochastic dynamic scheduling problem is very difficult or even impossible for large projects [4], mainly because the order of decisions is not fixed but depends on previous decisions and the realization of random variables [6].
Recognizing the shortcomings of model-based decision aids to create flexibility in plans, these are in practice replaced by judgmental decision processes [1, 3]. Dealing with the described complexity judgmentally is, however, not less complex, even when the large number of behavioral issues that may negatively affect the outcomes (limitations in working memory, incentive misalignment, invisible and illusory correlations, just to name a few) are “ignored.” In spite of this, the organization of shipbuilding projects at VARD demonstrates a high level of responsiveness and innovative solutions to quickly adapt technical changes throughout the construction processes.
Initial contextual studies point to “team abilities to share experience” and “tacit knowledge” as the drivers of responsiveness to changes. In social sciences, human relations and the value these relations have in sharing information, knowledge and resources define the social capital, which is quite often associated with enhanced innovation. Requesting and retrieving task related information and resources between project participants is considered a fundamental aspect of project execution [7].
Driven by the outlined challenges and industry practices, the design problem in planning complex projects was approached by connecting the classical O.R. elements of project scheduling with the social-behavioral aspects of handling technical uncertainty. In the case of VARD, this is being done through a long-term research engagement with SINTEF, the largest independent research organization in Scandinavia.
The multi-method approach extends the scope of research on classical O.R. approaches to project planning and scheduling in two distinct directions:
A. To connect the element of design to project scheduling, in one stochastic-dynamic program, to study the impact of design uncertainty on planning. Without this knowledge, it is difficult to achieve good solutions when advanced design is taking place concurrently with production.
B. To provide insight into how the behavioral characteristics of project participants in social networks – e.g., informal work connections, trust and risk behavior – affect responsiveness to unforeseen changes. This task connects social network perspectives to behavioral operations management.
The research steps with preliminary results and implementation are described below and outlined by Figure 3.
A. The O.R. approach – stochastic-dynamic programming
The planning complexity with stochastic changes in design specifications is approached by small model instances, with the aim to learn what it is that makes solutions good. This enables finding good plans in the future without actually solving complex stochastic models. The model is described in [9]. The results indicate high value of flexible hedging strategies (proactive strategies) that capture the value of future design changes.
The research demonstrates more than 35 percent cost reduction when applying proactive strategies compared to reactive strategies, where static (deterministic) plans are updated in light of new (design) information. This is exemplified by Figure 1 for increasing probability of design alternative B (i.e., decreasing belief in the originally assumed design alternative A). These results have great managerial value, as they indicate the cost-saving potential of proactive approaches to design uncertainty in planning, and the properties of these strategies with guidelines on when and where to develop flexibility and buffers in plans. The insights and planning guidelines developed, listed in [9], are valuable to improve judgmental decision-making. Simulation-based planning approaches may also benefit from the knowledge on what solutions structures should be investigated; particularly, when a full scenario tree evaluation is not possible due to the size of the problem.
B. The social-behavioral approach – Behavioral mechanisms in the social network of project work
Without insight into the social mechanisms, the integration of human behavior into planning and decision-making would be limited to understanding the processes and heuristics involved in individual level decision-making, as described by e.g. [5], which may differ from those driven by social interactions. Social ties are particularly critical when tacit knowledge constitutes a large share of the available resources as it is in many organizations with long traditions. While the transfer of explicit resources may (but need not to) follow formal process charts, effective transfer of tacit knowledge requires regular personal interaction and trust.
From a social network perspective, human interaction in a social environment can be expressed as a relation-based pattern, defined by symmetric or asymmetric (one-sided or two-sided) links between nodes (individuals) in the social environment. Such links or relations affect, for example, resource and information sharing, bargaining power and decision-making. Our social network analysis is, therefore, developed to provide insight into how the “informal” multiple relations (in daily work, technical uncertainty handling, friendship and true project network), trust and norms of behavior affect engineering responsiveness. The relational patterns are visualized, hypothesized and tested statistically [8].
The dynamic network visualization technique allowed quick identification of structural concerns in the engineering network; e.g., central and peripheral actors, strong triads that may facilitate or hinder the flow of information, gatekeepers and information brokers. These patterns exhibit different influences, behaviors and choices. For example, high centrality provides better access to critical information and theoretically is positively related to performance. Figure 2 exemplifies the “broker” role of one central actor in the technical uncertainty handling network. This actor connects the otherwise disconnected heterogeneous groups of project managers and technical coordinators (the two key functions in a project organization).
Figure 2: The “broker” role exemplified.
The social-behavioral approach uncovered high-performing groups and their ability to achieve synergies to handle technical changes in shipbuilding projects. These synergies are associated with the social capital of the group, mainly through cohesiveness of strongly interrelated actors, and through a positive and significant association between friendship and technical uncertainty handling networks. The solid base of interpersonal trust is translated into higher propensity for risk seeking [2]. These aspects, connected to extensive grade of tacit knowledge, in a flat organization that delegates decisions to lower levels, facilitates the development of high-performance teams, where people feel collaborative (positive) pressure to find immediate solutions to unplanned changes.
Industrial Implementation
The uncovered social-behavioral characteristics confirm the strong handcraft traditions in the shipbuilding project work, but also highlight the improvement potential by industrialization through craftsmanship. Such an industrialization process is the scope of the current change program at VARD, with a focus on reorganizing the engineering department to better align with the existing social capital while creating lean processes. In this change program, both the O.R. and social network-related findings are applied. One concrete action relates to front-end loading (pre-project planning), connecting engineering activities and planning to design and procurement early in a project’s lifecycle, at a time when the ability to influence is relatively high and the cost to make changes is relatively low. The aim is to early identify project uncertainty and activities with high impact on the schedule and performance, and develop proactive strategies to handle upcoming changes. The first ship with restructured engineering work is under construction, and great benefits are expected.
Figure 3: The multi-method approach to design uncertainty in shipbuilding planning.
Conclusion
The NextShip project results are not a single tool or model, but rather a collection of (and cautious use of) various tools, approaches and insights, all focused on handling the design problem in planning to improve project flexibility and responsiveness.
The results trigger increased focus on the social interactions and behavioral reactions to handle technical uncertainty in typical engineering environments. These changes are difficult to quantify and therefore difficult to manage. The reaction pattern through team dynamics compensates (to some extent) for lack of flexibility in plans, as people feel collaborative (positive) pressure to find immediate solutions to unplanned changes by making use of their social capital.
The role of O.R. cannot be ignored, though, as without analytical guidelines, a new improvement idea might fail. O.R. has a lot to offer in the planning of complex engineering projects, but its value is not necessarily in solving the complexity for large real projects, as this may be very difficult or even impossible. Small model instances of the problem have proven useful in developing understanding on what makes solutions good and to enable developing good plans without actually solving complex models. This potentially improves the outcome of judgmental decision processes.
However, because there is a “but” when uncertainty-handling is largely based on social interactions, it is not enough to know the right answer to what makes solutions good. It is also important to know how to utilize the power of social networks to transfer this knowledge.
The main message of this project is, therefore, on the high value of combining multiple research methodologies and disciplines to explore the real complexity of typical O.R. problems.
Hajnalka Vaagen is a senior scientist with the SINTEF Technology and Society, Dept. Applied Economics and Operations Research, headquartered in Trondheim, Norway.
Acknowledgment:
The research presented in this article is part of the NextShip project under Norwegian Research Council grant agreement 216418/O70. Special thanks to Arvid Sundli, senior vice president of operations, VARD Group AS, and Jan Emblemsvåg, senior vice president, Ship Design & Systems, Commercial Marine, for valuable input.
References
- Bendoly, E., Donohue, K., Schultz, K.L., 2006, “Behavior in operations management: Assessing recent findings and revisiting old assumptions,” Journal of Operations Management, Vol. 24, No. 6, pp. 737-752.
- Burt, R.S., 2005, “The Social Capital of Structural Holes” in Mauro F. Guillen and Randall Collins (Ed.), “The New Economic Sociology: Developments in an Emerging Field,” The Russell Sage Foundation.
- Emblemsvåg, J., 2014, “Lean project planning in shipbuilding,” Journal of Ship Production and Design, Vol. 30, No. 2, pp. 79-88.
- Jørgensen, T., Wallace, S. W., 2000, “Improving project cost estimation by taking into account managerial flexibility,” European Journal of Operational Research, Vol. 127, pp. 239-251.
- Kahneman, D., Tversky, A., 2000, “Choices, Values and Frames,” Cambridge University Press.
- Kall, P., Wallace, S. W., 1994, “Stochastic Programming,” John Wiley & Sons, Chichester.
- Katzenbach, J. R., Smith, D. K., 1993, “The wisdom of teams: Creating the high-performance organization,” Harvard Business Press.
- Vaagen H., Hansson, M, 2015, “The social-behavioural drivers of responsiveness in engineer-to-order projects,” under review.
- Vaagen H., Kaut M., Wallace S.W., 2015, “The impact of design uncertainty in engineer-to-order project planning,” under review.