Modeling the co-evolution of committee formation and awareness networks in organizations

Abstract

Large-scale organizations confront a difficult problem of assigning personnel to non-routine tasks. Such projects require novel combinations of individuals from separate parts of the organization. We propose a computational method for assembling a committee that recruits individuals to solve the task. This method is driven by data on interpersonal awareness relationships, fostered by mutual interaction and indirect contact. By inducing a network from these relationships, we can reduce the committee selection problem to the maximum coverage problem. A stochastic model is then presented to capture plausible new relationships which form during committee deliberation. This framework provides administrators the opportunity to track the evolution of awareness relationships while improving the organization’s ability to solve non-routine tasks. To track awareness network evolution, we demonstrate correlations in empirical networks between communication network topology and the probability of awareness relationships. To formalize organizational improvement over a series of committee formations, we extend a static committee formation problem to a repeated committee formation problem. After showing the computational cost of exact solutions to this problem, we propose an efficient, heuristic-based solution. Through simulation on real-world networks, we demonstrate that our approach produces more efficient and productive network states than a baseline algorithm during a series of committee formations.