Chapter 2 – Sensing and architecture of an agricultural cyber-physical system
Tommaso Adamo, Lucio Colizzi, Emanuela Guerriero, Shimon Y. Nof, Puwadol Oak Dusadeerungsikul
Abstract
This chapter explores the foundations of cyber-physical systems (CPS) and digital twins (DTs) applied to agriculture. It introduces a layered architecture composed of Internet of Things (IoT) nodes, backend services, databases, and application interfaces, emphasizing the complementarity between CPS and DT concepts. Core communication paradigms, publish–subscribe, push–pull, exclusive pair, and request–response, are presented and compared, showing their relevance in agricultural contexts. A didactic case study of a greenhouse with six irrigation lines illustrates the design of IoT sensor and actuator nodes, middleware, decision support systems, and a 3D DT implemented in FlexSim. The chapter is addressed both to domain experts, who need to understand the inherent complexity of such systems, and to developers, who can find practical insights for implementation. It concludes by highlighting the bidirectional nature of DTs, where physical events are mirrored in the virtual model and user actions in the digital space affect the real environment.
https://doi.org/10.1016/B978-0-443-43918-6.00007-6
Chapter 4 – Decision-making in smart agriculture: Advanced models and techniques
Tommaso Adamo, Lucio Colizzi, Emanuela Guerriero
Abstract
This chapter addresses one of the core challenges in smart agriculture: how to support complex, real-time decision-making processes through advanced computational paradigms. As agriculture evolves into a data-intensive domain, driven by artificial intelligence, Internet of Things, and cloud-based analytics, traditional decision support tools are no longer sufficient to manage the increasing complexity and dynamism of agronomic systems. Smart farming demands models that can capture biodiversity optimization, real-time fertigation strategies, and efficient use of limited resources. We illustrate a robust solution approach based on a model-and-run paradigm, centered on the separation between problem modeling and problem solving. We distinguish between two prominent model-and-run frameworks: Constraint programming (CP) and mathematical programming (MP), comparing their strengths, inference methods, and application domains. The chapter is structured around two real-world smart farming decision problems. Section 4.2 focuses on optimizing crop planting layout to maximize biodiversity using CP. Section 4.3 tackles fertigation and irrigation scheduling under a data-driven, sequential decision-making perspective using integer linear programming (ILP). Each section outlines the decision problem, provides its formal encoding and notation, and develops the computational model. To demonstrate the practical relevance of these models, both approaches have been deployed within a microservices-based architecture composed of a database, a front-end application, and two back-end services, invoking respectively a CP solver and an ILP solver. This architecture exemplifies how decision modeling techniques can transition from research to operational deployment, enabling next-generation digital agriculture systems.
Leave a Reply