We will discuss what we mean by managerial decision making, with examples from business, environmental, and societal domains. The challenges of effective decision making will be related to control concepts. “Control science and engineering is the only rigorous approach to effective decision-making in complex dynamical systems”—the relevance of this mantra will be emphasized. We will also note that decision-making can be seen as model-based control, with “mental models” taking the role of mathematical models where the latter are infeasible. This analogy suggests rich avenues for exploring the connections between human decision-making and control. For example, rigorous methods from control science for dealing with model uncertainty bear upon the use of mental models in organizations.

Systems thinking, a discipline that emerged from control engineering, is widely used in business and societal domains. This talk will provide a brief historical perspective on the topic, including connections with system dynamics (causal dynamic models). The concept of mental models will be discussed in connection with the the work of Kahneman and Tversky, which demonstrates the biases and illusions that human decision-makers are subject to. The talk will explore the extent to which modern management decision making theory and practice reflects important concepts from control theory and where opportunities exist to enhance current practice given today's societal demands and imperatives.

Applying Control Concepts to Supply Chain Decision-Making – Bhagyesh Patil. This talk will present two case studies of applying control concepts to decision making problems in the supply chain domain, discussing the role of control elements such as feedback, time-delay, and model adaptiveness. The first case study is on supplier selection: a decision-making process to select suppliers based on factors like price, quality, delivery time, minimum ordering, and reliability. How control principles are leveraged in this decision-making will be then presented through a simple real-world example. Second, we will discuss raw inventory ordering:  a multi-faceted optimization problem to order parts for manufacturing processes. We will show through a real-world example how control and human-in-the-loop principles are involved, with reference to inventory costs, carrying costs, and relationships with suppliers.

Human-machine interaction is becoming crucial for enhancing robustness in decision making in complex production environments, particularly under environmental uncertainties like in the process industries. We will explore a methodology for analyzing AI/ML and human collaboration within the MAPE- K (Monitor-Analyze-Plan-Execute over a shared Knowledge) framework, exploring its application in diverse process industries such as asphalt, steel, pharmaceuticals, and aluminum production. In addition, we will also discuss how digital transformation is enabling circular economies and new business models, where decision making, data sovereignty, and collaboration between entities are governed in a closed loop so as to create a trustful work opportunity. 

(This talk will be presented as a prerecorded video with Q&A over Zoom.)

What can feedback principles, so often based on rigid mathematical analysis, provide to systems for which any mathematical rigor is hard to find? Our approach is to think of fundamental feedback principles as guidelines, rather than actual rules. We believe those guidelines provide a rich source of correction for business processes and policymaking. In the end our feedback-fundamentals-inspired guidelines may not guarantee us only correct decisions, but they can keep us away from harmful practices. We will also consider the addition of feedfoward as a metaphor for planning in business and decision-making processes.  As with feedforward control methods, business planning depends entirely upon the quality of the models used, but even when these models are excellent, they cannot account for outside events. Using the example of industrial and infrastructure policy, we can also compare the effects of imperfect planning versus no planning and see how the lessons of combined feedfoward/feedback methods can provide some intuition.

A set of 10-12 “insights” from control science and engineering will be presented and discussed, in each case emphasizing the relevance for managerial decision-making. The insights will reprise earlier discussions in the workshop—such as feedback/feedforward integration and the mathematical- vs. mental-model analogy—as well as new examples—such as robustness/adaptation/performance interplays, the importance of state estimation, uncertainty management, and the decision-making challenges arising from delays in feedback systems.