This paper investigates how political decisions emerge from

complex, multi-actor governance environments using an agent-based modelling (ABM) framework. Building on a multi-layered urban climate governance model developed for the municipality of Askøy, near askøy, Norway, we isolate and systematically analyze the political decision-making component of the process. The model integrates signals from urban planners, societal feedback, non-governmental organizations (NGOs), and

media, which are aggregated by heterogeneous political agents through a two-stage process combining utility-based evaluation and peer influence dynamics. Through controlled simulation experiments, we examine how decision outcomes (acceptance, rejection, or revision) depend on factors such as signal configurations, decision thresholds, and party affiliation. Results reveal that decision dynamics are highly non-linear, characterized by tipping points and regime shifts rather than simple additive effects of inputs. While external signals largely determine outcomes under aligned

conditions, internal political factors (such as party alignment, individual preferences, and strategic orientation) become critical in scenarios with high conflict. The findings provide a structured understanding of the aggregation mechanisms linking various input signals to political outcomes.

Opinion dynamics play an important role in a wide variety of contexts and topics. While opinion formation is largely based on memories of previous interactions and opinion similarities with one's social network, existing social simulation models are often still rule based, and do not address cognitive functions such as memory decay and avoiding asking the same question to the same other. This paper investigates how augmenting agents with simulated human memory, based on ACT-R's declarative memory system \cite{Anderson2004}, will impact opinion dynamic simulations. Therefore, the micro- and macro behaviour of two HUMAT agent-based models \cite{jager2025humat}, one with and one without ACT-R declarative memory, were compared. The results demonstrate that including a declarative memory produces more realistic dynamics, and has an impact on the opinion at the scale of the population, indicating that opinion dynamics models can benefit from including a declarative memory.

This paper introduces an agent-based modeling framework that simulates con-sumer choice based on human need satisfaction and social comparison dynamics. The behavioral rules aim to capture the evolving interplay between individual well-being and social influences on consumption. The current version of the model focuses on replicating empirical consumption patterns and stylized facts, serving as a validation step for the underlying behavioral assumptions for future model applications.