Abstract
PurposeThis study aims to specify whether heterogeneous reference-point-based aspirations are related to the cooperation levels of R&D alliance portfolios in a positive or negative (or nonlinear) way, and to unveil how cooperative behaviors evolve in recurrent project cooperation.Design/methodology/approachThis study establishes a network containing a cooperation subnetwork and a project subnetwork based on patent data in the “deep learning” field to investigate how cooperative behaviors evolve in R&D alliance portfolios. A model of evolutionary games on complex networks is constructed to gain insight into the dynamic evolution of DMs’ strategies.FindingsFirst, the heterogeneous aspirations of DMs can improve the cooperation level in R&D alliance portfolios. Second, compared to prudent DMs, aggressive DMs are more likely to choose the cooperation strategy, implying that an appropriate aspiration level nurtures cooperative R&D endeavors with partners. Third, the effects of effort complementarity, knowledge reorganization capabilities and cooperation supervision on cooperation are contingent on the distribution of DMs’ aspiration types.Practical implicationsPolicymakers should identify aspiration types of DMs when screening partners. They can encourage partners to focus more on historical payoffs and establish relatively higher aspiration levels to improve the cooperation level. Developing highly detailed contracts becomes crucial when cooperating with firms that possess extensive knowledge reorganization capabilities.Originality/valueThis work contributes a theoretical framework for investigating cooperation in R&D alliance portfolios through the lens of evolutionary games on complex networks, thus revealing the effects of heterogeneous reference-point-based aspirations of DMs on R&D cooperation.