This paper focuses on the internal representation of norms and the necessary concepts required for normative decisionmaking in computational agents. We leverage the social science literature to integrate currently absent prevalent concepts of social expectations within the Expectation Event Calculus (EEC) and, in doing so, extend the formalism. Through the adopted terminology of expectations, we distinguish between descriptive and social norms, enabling a more comprehensive description of individual and collective behavior conditional on social expectations. We introduce complementary abstractions for normative attributes to demonstrate and explain why such a distinction between expectations enables richer normative scenarios to be modeled, which has yet to be shown in the EEC. We demonstrate this extension through a binarydecision social scenario. First, through a single-agent implementation driven by hardwired narratives. Secondly, we demonstrate the extension through a selection of multi-agent scenarios that showcase a change in behavior conditional on expectations. As a minimal implementation of social expectations, we conclude the paper with themes and open challenges as avenues for further research.

Cooperation has been widely studied in multi-agent foraging tasks. However, the impact of agent-environment interactions on the longer term and the achievement of sustainability have been largely unexplored in this context. This work contributes to the development of a testbed for exploring social dynamics between agents: the ‘sustainable foraging problem’. This testbed explores the effect of agent behaviour and the agent’s dilemma of choosing between individual reward and collective long-term goals for sustainable resource management. To incorporate varied levels of replenishment rates in this testbed, forest, pasture and desert environment types are formulated. A co-evolving deliberative loop with neuro-evolution that asks the agents to act with greedy or moderate behaviour is demonstrated. This deliberative layer is shown to be insufficient in situations of social dilemma where the agents learn to increase their individual rewards instead of collectively increasing these rewards through the sustainability of the environment. A simple reflective governor based on the notion of the agent’s self-awareness is illustrated to allow the agents to occasionally reason about the long-term impacts of their immediate actions on future resource availability in the environment, which may eventually ensure sustainability.

Evolving agents to learn how to solve complex, multi-stage tasks to achieve a goal is a challenging problem. Problems such as the River Crossing Task are used to explore how these agents evolve and what they learn, but it is still often difficult to explain why agents behave in the way they do. We present the Minimal River Crossing (RC-) Task testbed, designed to reduce the complexity of the original River Crossing Task while keeping its essential components, such that the fundamental learning challenges it presents can be understood in more detail. Specifically to illustrate this, we demonstrate that the RC- environment can be used to investigate the effect that a cost to movement has on agent evolution and learning, and more importantly that the findings obtained as a result can be generalised back to the original River Crossing Task.

Taking inspiration from nature for meta-heuristics has proven popular and relatively successful. Many are inspired by the collective intelligence exhibited by insects, fish and birds. However, there is a question over their scalability to the types of complex problems experienced in the modern world. Natural systems evolved to solve simpler problems effectively, replicating these processes for complex problems may suffer from inefficiencies. Several causal factors can impact scalability; computational complexity, memory requirements or pure problem intractability. Supporting evidence is provided using a case study in Ant Colony Optimisation (ACO) regards tackling increasingly complex real-world fleet optimisation problems. This paper hypothesizes that contrary to common intuition, bio-inspired collective intelligence techniques by their very nature exhibit poor scalability in cases of high dimensionality when large degrees of decision making are required. Facilitating scaling of bio-inspired algorithms necessitates reducing this decision making. To support this hypothesis, an enhanced Partial-ACO technique is presented which effectively reduces ant decision making. Reducing the decision making required by ants by up to 90% results in markedly improved effectiveness and reduced runtimes for increasingly complex fleet optimisation problems. Reductions in traversal timings of 40–50% are achieved for problems with up to 45 vehicles and 437 jobs.