Automation, Research and Investment Policies in Firms

Abstract

In our time automation, in combination with burgeoning fields such as Artificial Intelligence, has grown to be a significant factor, and with it the role of scientific and engineering knowledge in the working of firms has grown too. We present a model-based study of firms in which generation of new knowledge, and the application of accumulated knowledge are integral to business since these determine the range and scope of the firms' products, and also the efficiency of R&D and the production process. In our model the firm is organized as two functionally separate stages in series. Stage 1's activity is R&D which creates new concepts, methods and prototypes of products. Stage 2 deploys labor and capital, in the form of "machines'', in production, which transforms selected outputs of Stage 1 into marketable, profitable products. New knowledge is generated from dedicated research in Stage 1 as well as by Learning-by-Doing (LbD) in both stages. Knowledge is subject to obsolescence over time. The firm's investment policy determines the allocation of funds to each stage subject to a budget constraint, operations management controls the admission of the output of Stage 1 to Stage 2, and also the combination of labor and machines in Stage 2. We analyze the interaction of these decisions, and the dynamical evolution of the knowledge stock under two management strategies. The short-term-focused, myopic strategy takes the existing knowledge stock as given, and maximizes the immediate profit. The long-term-focused strategy takes into account the future benefits of generating new knowledge in the investment decision. We use commonly-used production functions to obtain nonlinear dynamical system models, which are analyzed. We show that for both strategies the system converges to a steady-state where the knowledge stock and investment allocation remain constant over time. In numerical studies we compare the system behavior for the two strategies, and characterize their dependencies on various factors, such as the strength of the LbD effect, return on knowledge stock, and the influence of knowledge in expanding the scope and range of the firm's products.