Here we discuss technology development models available in literature and (hypotheses on) the relationship of supraregional research collaboration and input acquisition with technological progress in regions.
Technology Development Models
To use an agent-based model to study the effects of supraregional relationships on technological progress, one requires a model of technology development. In the economics of technological change, there is a small body of literature that provides fundamental, operational models of technology development. In general, in the models discussed here, technology is perceived as a collection of elements and the technology’s feasibility or fitness depends on how these elements are structured (e.g. ordered). In these models, there is a search heuristic changing (1) the set of elements to be combined, or (2) the structure of these elements. In Padgett et al. (2003), a technology is modeled as a series of elements (chemical processes), where the technology is feasible whenever a combination of these elements has a particular order (the chemical processes form a ‘hypercycle’). Technology search concerns randomly combining given elements. In Korhonen and Kasmire (2013), a technology is a series of elements (numbers) that are interconnected by transformations (arithmetic operators). Technology is feasible if the arithmetic outcome meets a predefined number. Technology search concerns finding a combination of elements and transformations. In Arthur and Polak (2006), a technology is constructed by trial-and-error combination of elements (logical circuits) into a system that is feasible if it meets a certain predefined outcome (logical input- output requirement). In search, previously discovered systems can be used as building blocks. In Frenken and Nuvolari (2003), the technological structure is given, but each element may have two or more options, each with an own fitness contribution that also depends on the options chosen for the other elements. Technology search concerns finding the most ‘fit’ combination of options. In Gilbert et al. (2001), technology is a combination of elements (units of technological knowledge) drawn from an agent’s subset of the universe of elements. Each combination has a certain (product payoff). Technology search concerns finding higher payoff by (1) changing the subset and (2) changing the combination. In Silverberg and Verspagen (2005), technologies are cells on a cylinder and they are feasible only when they are adjacent to at least one feasible technology or the bottom of the cylinder. Technology search concerns ‘discovering’ and possibly making feasible of new cells. In Morone and Taylor (2010), technologies are nodes in a predefined, fixed directed graph. While technologies can be discovered by randomly traversing the graph, they become feasible only when all of the ‘ancestors’ are discovered and feasible.
As we study the role of both research collaboration and input acquisition, we need to disaggregate the conception of technology. For this, we turn to the field of Operations Management, in which firms work with a Manufacturing Bill-of-Mate- rial (MBOM). Such a bill-of-material is the ‘recipe’ that specifies (1) the inputs, intermediate products, and raw resources required to make an output product, and (2) all operations and the sequence of manufacturing steps to transform input, intermediate products, and raw resources into one or multiple output products. In our model, we disaggregate the technology search space by discerning search for concrete artifacts and search for transformations (‘means’) required to produce those artifact. Chie and Chen (2013) also model products as a bimodal tree of transformations and artifacts.
Some of the technology development models have explicit ‘economic agents’ (e.g. firms) that perform the search. Since technical elements are combined into technology in these models, collaboration is (or: may be) modeled by having agents contribute a subset of elements to that jointly created technology. An additional benefit is (or: may be) that agents thus obtain technical elements from collaborators (Gilbert et al. 2001; Morone and Taylor 2010). On top of this, Morone and Taylor (2010) provide a technology development model with a spatial dimension in the sense that agents only exchange technological elements when they are located close to each other. In our model, firm agents within reach of each other may collaborate in two ways: by providing input artifacts to be used by other agents and by combining transformations to ‘unlock’ new transformations. Both will be explained in great detail later.