Criminal organizations operate in complex changing environments. Being flexible and dynamic allows criminal networks not only to exploit new illicit opportunities but also to react to law enforcement attempts at disruption, enhancing the persistence of these networks over time. Most studies investigating network disruption have examined organizational structures before and after the arrests of some actors but have disregarded groups’ adaptation strategies.
MADTOR simulates drug trafficking and dealing activities by organized criminal groups and their reactions to law enforcement attempts at disruption. The simulation relied on information retrieved from a detailed court order against a large-scale Italian drug trafficking organization (DTO) and from the literature.
The results showed that the higher the proportion of members arrested, the greater the challenges for DTOs, with higher rates of disrupted organizations and long-term consequences for surviving DTOs. Second, targeting members performing specific tasks had different impacts on DTO resilience: targeting traffickers resulted in the highest rates of DTO disruption, while targeting actors in charge of more redundant tasks (e.g., retailers) had smaller but significant impacts. Third, the model examined the resistance and resilience of DTOs adopting different strategies in the security/efficiency trade-off. Efficient DTOs were more resilient, outperforming secure DTOs in terms of reactions to a single, equal attempt at disruption. Conversely, secure DTOs were more resistant, displaying higher survival rates than efficient DTOs when considering the differentiated frequency and effectiveness of law enforcement interventions on DTOs having different focuses in the security/efficiency trade-off.
Overall, the model demonstrated that law enforcement interventions are often critical events for DTOs, with high rates of both first intention (i.e., DTOs directly disrupted by the intervention) and second intention (i.e., DTOs terminating their activities due to the unsustainability of the intervention’s short-term consequences) culminating in dismantlement. However, surviving DTOs always displayed a high level of resilience, with effective strategies in place to react to threatening events and to continue drug trafficking and dealing.

This multi-model (i.e. a model composed of interacting submodels) is a multi-level representation of a collective motion phenomenon. It was designed to study the impact of the mutual influences between individuals and groups in collective motion.

The model explores how two types of information - social (in the form of pheromone trails) and private (in the form of route memories) affect ant colony level foraging in a variable enviroment.

The CHIME ABM explores information distribution networks and agents’ protective decision making in the context of hurricane landfall.

RAGE models a stylized common property grazing system. Agents follow a certain behavioral type. The model allows analyzing how household behavior with respect to a social norm on pasture resting affects long-term social-ecological system dynamics.

Large-scale land acquisitions (LSLAs) threaten smallholder livelihoods globally. Despite more than a decade of research on the LSLA phenomenon, it remains a challenge to identify governance conditions that may foster beneficial outcomes for both smallholders and investors. One potentially promising strategy toward this end is contract farming (CF), which more directly involves smallholder households in commodity production than conditions of acquisition and displacement.

To improve understanding of how CF may mediate the outcomes of LSLAs, we developed an agent-based model of smallholder livelihoods, which we used as a virtual laboratory to experiment on a range of hypothetical LSLA and CF implementation scenarios.

The model represents a community of smallholder households in a mixed crop-livestock system. Each agent farms their own land and manages a herd of livestock. Agents can also engage in off-farm employment, for which they earn a fixed wage and compete for a limited number of jobs. The principal model outputs include measures of household food security (representing access to a single, staple food crop) and agricultural production (of a single, staple food crop).

…

A road freight transport (RFT) operation involves the participation of several types of companies in its execution. The TRANSOPE model simulates the subcontracting process between 3 types of companies: Freight Forwarders (FF), Transport Companies (TC) and self-employed carriers (CA). These companies (agents) form transport outsourcing chains (TOCs) by making decisions based on supplier selection criteria and transaction acceptance criteria. Through their participation in TOCs, companies are able to learn and exchange information, so that knowledge becomes another important factor in new collaborations. The model can replicate multiple subcontracting situations at a local and regional geographic level.
The succession of n operations over d days provides two types of results: 1) Social Complex Networks, and 2) Spatial knowledge accumulation environments. The combination of these results is used to identify the emergence of new logistics clusters. The types of actors involved as well as the variables and parameters used have their justification in a survey of transport experts and in the existing literature on the subject.
As a result of a preferential selection process, the distribution of activity among agents shows to be highly uneven. The cumulative network resulting from the self-organisation of the system suggests a structure similar to scale-free networks (Albert & Barabási, 2001). In this sense, new agents join the network according to the needs of the market. Similarly, the network of preferential relationships persists over time. Here, knowledge transfer plays a key role in the assignment of central connector roles, whose participation in the outsourcing network is even more decisive in situations of scarcity of transport contracts.

This is an extension of the original RAGE model (Dressler et al. 2018), where we add learning capabilities to agents, specifically learning-by-doing and social learning (two processes central to adaptive (co-)management).

The extension module is applied to smallholder farmers’ decision-making - here, a pasture (patch) is the private property of the household (agent) placed on it and there is no movement of the households. Households observe the state of the pasture and their neighrbours to make decisions on how many livestock to place on their pasture every year. Three new behavioural types are created (which cannot be combined with the original ones): E-RO (baseline behaviour), E-LBD (learning-by-doing) and E-RO-SL1 (social learning). Similarly to the original model, these three types can be compared regarding long-term social-ecological performance. In addition, a global strategy switching option (corresponding to double-loop learning) allows users to study how behavioural strategies diffuse in a heterogeneous population of learning and non-learning agents.

An important modification of the original model is that extension agents are heterogeneous in how they deal with uncertainty. This is represented by an agent property, called the r-parameter (household-risk-att in the code). The r-parameter is catch-all for various factors that form an agent’s disposition to act in a certain way, such as: uncertainty in the sensing (partial observability of the resource system), noise in the information received, or an inherent characteristic of the agent, for instance, their risk attitude.

The wisdom of the crowd refers to the phenomenon in which a group of individuals, each making independent decisions, can collectively arrive at highly accurate solutions—often more accurate than any individual within the group. This principle relies heavily on independence: if individual opinions are unbiased and uncorrelated, their errors tend to cancel out when averaged, reducing overall bias. However, in real-world social networks, individuals are often influenced by their neighbors, introducing correlations between decisions. Such social influence can amplify biases, disrupting the benefits of independent voting. This trade-off between independence and interdependence has striking parallels to ensemble learning methods in machine learning. Bagging (bootstrap aggregating) improves classification performance by combining independently trained weak learners, reducing bias. Boosting, on the other hand, explicitly introduces sequential dependence among learners, where each learner focuses on correcting the errors of its predecessors. This process can reinforce biases present in the data even if it reduces variance. Here, we introduce a new meta-algorithm, casting, which captures this biological and computational trade-off. Casting forms partially connected groups (“castes”) of weak learners that are internally linked through boosting, while the castes themselves remain independent and are aggregated using bagging. This creates a continuum between full independence (i.e., bagging) and full dependence (i.e., boosting). This method allows for the testing of model capabilities across values of the hyperparameter which controls connectedness. We specifically investigate classification tasks, but the method can be used for regression tasks as well. Ultimately, casting can provide insights for how real systems contend with classification problems.

The agent based model matches origins and destinations using employment search methods at the individual level.