On the sustainability of complex systems : understanding the effects of disruptions in the sustainability of supply networks

A supply network (SN) can be broadly defined as a system with parties involved in fulfilling costumer’s orders, where the essential principle is that each of them aims to maximise its profit or utility. These systems are now more complex and intertwined than ever, meaning that their modelling and assessment should include aspects of their complexity. Effects resulting from the affectation of nodes, or the SN adaptability against disruptive events hardly follow a linear fashion
and can only be identified when system’s evolution or companies’ agency are included in the SN model. The understanding of disruptive effects, in particular, is a relevant topic since recent events have demonstrated that perturbations on SNs can propagate and generate damages to the environment and the social dimensions. While the literature has shown efforts to include complexity into methods when proposing disruption mitigation approaches, similar attempts 3in the field of sustainability assessment are scarce, leading to the overlook of this intrinsic SN characteristic. We argue that the strategies meant to allow the achievement of society goals, such as sustainable development, should contemplate these effects since they model systems that can show complex behaviors. We focused on determining how could we incorporate complexity characteristics into the sustainability assessment of SNs, having the study of disruptive events as an example.

We split the body of work into three parts. In the first part, we conducted a literature review in which we identified the conceptual differences between sustainability and resilience from an epistemological perspective. We identified that these concepts are decoupled from a methodological, motivational, and temporal perspective. Using these findings, we elaborated a set of four principles that should guide the proposal of any complexity-driven sustainability assessment approach. This leads to our first contribution, which is a sustainability assessment framework that is underpinned in the four principles, and that uses agent-based modelling (ABM) as main modelling paradigm.

In the second part, we dealt with the lack of flexibility and replicability of current ABM approaches that diminishes its usefulness when solving sustainability-related inquires. For this, we proposed AFRICA, a mathematical framework rooted on principles of algebraic life cycle assessment, designed to represent socio-technical agents in agent-based simulations. In addition to this, we presented the software pacha, which can perform agent-based simulations and it is designed to fit the AFRICA framework. The contribution of this part is the AFRICA framework itself because it is mathematical, flexible, and language-agnostic, so it can be implemented in any programming language.

In part three, we put to test the two previous contributions by proposing two cases of study. In the first case, we studied the effects of introducing agents with sustainable attitudes, also named Agents of Change, in a supply network. In the second case, we studied the effects that the introduction of disruptive events can have on the sustainability of the Peruvian fishmeal industry. For the first case,
we demonstrated that there exist strategies and network configurations where the adoption of an environmental friendly business norm can reduce considerably the impact of the system whilst not representing a riskier decision from a financial standpoint. In the second case of study, we showed that accumulation of impacts and its trend can vary disproportionally when unexpected disruptions
are introduced. Moreover, disruptions on periods of high productivity, can produce changes from which the system will hardly recover.

Finally, in this thesis, the show that the sustainability of systems under disruptive effects can be addressed by incorporating a complexity-oriented perspective into the modelling exercise. Moreover, thanks to the cases of studies, we identified two streams in which sustainability research can advance, and to which this thesis contributes. The first stream is oriented to the study of fundamental questions about the sustainability of systems, so theories, principles, or experiments can be proposed and studied. The second stream is focused on enhancing the current assessment approaches by introducing agency and dynamic components into the modelling and the sustainability assessment.