Industrial agents (IAs) [1] are multiagentbased systems (MASs) [2] that, for many years, have been advocated as a promising and realistic solution for an emerging set of industrial challenges. In the past, MASs fell into the scope of enterprise agility [3]–[8], and now, more than ever, pertain to the industrial digital transformation and sustainability spheres. MAS technology is being applied to several industrial applications in the cyber-physical system (CPS) context, namely, in smart production, smart electric grids, smart logistics, and smart health care [9]. To understand the future potential of IAs, one must first have a sufficiently concise view of the past and present efforts, i.e., understand their early applications and current directions. Such a view is necessary because, over the last decades, the concept of the IA has proven to be a bit of a moving target, adjusting to the needs, visions, and technologies of each era.
As an example, Figure 1 reflects this reality well by depicting the most prevalent words in IA research articles
[10] published by the IEEE Technical Committee on IAs [11], and it makes it evident that they encompass a wide range of industrial system aspects. This work presents a view of the authors and is driven by looking at past and recent efforts in the IA domain, current trends, and relevance, with key enabling features of industrial CPSs (ICPSs) [12], [13], to assess how the agent definitions, development, and solutions have changed over the years and where current efforts may focus.
