Advances in Complex Systems

There has been a surge of interest in community detection in homogeneous single-relational networks which contain only one type of nodes and edges. However, many real-world systems are naturally described as heterogeneous multi-relational networks which contain multiple types of nodes and edges. In this paper, we propose a new method for detecting communities in such networks. Our method is based on optimizing the composite modularity, which is a new modularity proposed for evaluating partitions of a heterogeneous multi-relational network into communities. Our method is parameter-free, scalable, and suitable for various networks with general structure. We demonstrate that it outperforms the state-of-the-art techniques in detecting pre-planted communities in synthetic networks. Applied to a real-world Digg network, it successfully detects meaningful communities.

Community detection is an important issue to understand the structural and functional properties of complex networks, which still remains a challenging subject. In some complex networks, a node may belong to multiple communities, implying overlapping community structure. Moreover, complex networks often show a hierarchical structure where small communities group together to form larger ones. In this paper, we propose a novel parameter-free algorithm called agglomerative clustering based on label propagation algorithm (ACLPA) to detect both overlapping and hierarchical community structure in complex networks. By combining the advantages of agglomerative clustering and label propagation, our algorithm can build the hierarchical tree of overlapping communities in large-scale networks. The tests on both synthetic and real-world networks give excellent results and demonstrate the effectiveness and efficiency of our algorithm.

This paper suggests an opinion dynamics approach to define community structures in complex networks. If a typical opinion dynamics model is applied to a network with a community structure, the network can separate in two groups of nodes. Such bisection in a given network can arise in many different ways depending on the initial conditions. The opinion distance between two nodes is defined as the probability of disagreement, or the probability that the two nodes belong to different bisections in multiple Monte Carlo simulations. The communities can be defined in terms of the distance. Closer nodes belong to the same community. Three opinion dynamics models were tested to show how the method works. Through various example networks, it was shown that the distance data can be used as a unique metric for identifying hierarchical structures and overlapping nodes in networks, as well as for identifying the community structure itself.

Mobility of humans plays an important role in shaping the performance of various services in mobile and location dependent networks. In this paper, we amalgamate and summarize the parameters reported in various trace analyses, and explore the inter-dependencies among the parameters using a layered approach. We then identify using an experimental setup (a) several roadblocks in generating an intended synthesis model, (b) the changes in the mobility patterns with variation in inputs (e.g., the underlying map), and (c) the changes in efficiency of a given service (here we have considered simple broadcasting) with change in the input parameters. Interestingly, we note that the efficiency of the service does not necessarily depend on the mobility pattern and one needs to be extremely careful before drawing a direct correlation between the two.

The primary and secondary effects of various spatial and transportation policies can be evaluated with models of activity–travel behavior. Whereas existing activity-based models of travel demand simulate a typical day, dynamic models simulate behavioral response to endogenous or exogenous change, along various time horizons. The current study aims at developing a model of endogenous dynamics of activity–travel behavior. Endogenous dynamics are induced by stress, which is regarded as dissatisfaction with current habits. It is assumed that people try to alleviate stress by trying short-term changes, within the options known to them or by exploring new options. If these explorations prove to be unsuccessful, they will consider long-term changes, such as moving to a new residential location, buying a car, etc. Therefore, this self-improvement process can result in both short and long-term adaptations. In the proposed framework, choice-set formation is modeled, the key concepts of aspiration, activation, awareness and expected utility are integrated, while both rational and emotional mechanisms are taken into account. Numerical simulations are conducted in order to check the face validity of the model, as well as the impact of stress tolerance parameters on system performance.