26th International Association for Management of Technology Conference - IAMOT 2017

Developing a risk-adaptive technology roadmap (TRM) through bayesian network and association rule mining under deep uncertainty Yujin Jeong and Byungun Yoon Firms always face rapidly changing and complex environment, which is carefully dealt with because it is directly connected with success and failure in business. Under this circumstance, shorter technology life cycle and higher complexity of technology cause risk and uncertainty. Thus, it is necessary to forecast future changes and respond in time, and a diversity of studies has generated to identify risk and uncertainty as well as to planning under deep uncertainty. In particular, scenario-based technology roadmapping has received much attention because scenario planning is able to cope with uncertain events systematically. But it requires much times and efforts and the results are not highly concrete, which fails to establish more accurate and specific strategy. In short, there is a lack of communication with strategy and it reduces the advantage of TRM which is a tool communicating strategy with technology, product and market. They make it hard to update and complement TRM in contexts while reducing sustainability of planning. Therefore, this paper aims to develop a risk-adaptive TRM which makes possible to update and renew TRM at risk based on various risk factor affecting technology and market innovation. At first, risk factors are identified through text mining and opinion mining from relevant databases within the framework of STEPPER (Society, technology, environment, population, politics, economy and resource). Then, possible states for each node are defined in the pre-developed TRM through developing the event tree which consists of trigger events caused by risk factor serving as basic events. Second, probability and impact of each state are calculated by Bayesian network which is able to deal with uncertain knowledge by using probability information and conjoint analysis that provides partial value of each attribute level as well as utility of each profile with multiple attributes. Especially, the partial value is used as impact of each state, and the utility by each profile which is the combination of node in TRM and results in the path for achieving strategic goal will be one of elements for evaluating outcome. Third, the basic action plan is constructed by combining each nodes and arcs with consideration of all possible states. Fourth, outcome of action plan is evaluated through creating new indicator – feasibility based upon impact, importance, utility and strategic fit. It serves as the criteria to determine whether the state for node is adapted or perished. Finally, the risk-adaptive TRM is completed by adjusting and reconstructing paths based on feasibility and severity of risk event at each signpost when supposing that future events related to risk factor are occurred at the specific point. The basic action plan is adapted by re-selecting each node and arc with the goal of increasing strategic fitness and utility for users, and it leads to transform basic action plans by reflecting risk factors and whether they occur or not. Consequently, the risk-adaptive TRM is able to overcome the limitations that traditional TRM underestimate complex environment surrounding innovations in technology and market. The proposed roadmap enables to increase success rate for implementing and commercializing new technology because it attempted to adapt or perish insignificant actions. In practice, the risk-adaptive TRM helps managers to recognize turbulent environment easily and make decisions by considering the results quantitatively measured on the basis of technological and market information.