Real-time strategy (RTS) games simulate battlefield leadership and tactical and strategic operations. Most overemphasize the number of actions per minute (APM), which encourages players to click rapidly and constantly rather than apply deliberate and finely tuned strategies or tactics. New RTS games featuring resource dispersion game mechanics aimed at reducing APM demand and promoting strategic planning are being released. We created three versions of a single RTS game, recruited players, recorded their game control data, and asked them to complete a simple after-game questionnaire. Data were used to analyze tactical and strategic applications. Player actions were observed and player opinions analyzed in an attempt to identify an optimal game structure as measured by strategic and tactical play.

Complex economic phenomena require significant time and effort to make observations and collect data. Inspired by the concepts of indicator species and the Big Mac index, in this paper we propose a novel and intuitive market concept, indicator products, for analyzing purchase decisions that are made on a regular basis, and for helping online game designers and researchers to observe player behaviors and market conditions. Blizzard Entertainment's World of Warcraft is used to show how indicator products can be identified. We discuss the common features that indicator products share, as well as associated player behaviors.

The identification of multiple network spreaders is an appropriate solution to spread information, ideas or diseases in many practical applications. For instance, in target marketing, the spreaders are selected from customer groups classified by similar purchase behaviors to advertise the products, and to optimize the allocation of limited resources. The community detection approaches intuitively are used to identify the community structures or social groups in a social/complex network. However, how to determine the number of community K is a difficult issue. Hence, two-phase evolutionary framework (TPEF) is proposed for automatically determining the number of community K and maximizing the modularity of communities. In the preliminary experiment, the LFR benchmark networks are used to test the proposed method, and to analyze the execution time, the community quality and the network spreading effect. The experiment results show that TPEF can perform well and produce the satisfied quality of community structures. The community detection approaches can be used to assist selecting the multiple network spreaders, and to gain the benefit in network spreading when the community structure is obvious. Furthermore, our results suggest that developing an index, a mechanism or a sampling technic is necessary to decide whether the community detection approaches are applied for selecting multiple network spreaders.

Almost all current matchmaking systems for team competition games based on player skill ratings contain algorithms designed to create teams consisting of players at similar skill levels. However, these systems overlook the important factor of playing style. In this paper, we analyze how playing style affects enjoyment in team competition games, using a mix of Sternberg's thinking style theory and individual histories in the form of statistics from previous matches to categorize League of Legend (LoL) players. Data for approximately 64 000 matches involving 185 000 players were taken from the LoLBase website. Match enjoyment was considered low when games lasted for 26 min or less (the earliest possible surrender time). Results from statistical analyses indicate that players with certain playing styles were more likely to enhance both game enjoyment and team strength. We also used a neural network model to test the usefulness of playing style information in predicting match quality. It is our hope that these results will support the establishment of more efficient matchmaking systems.

The identification of influential spreaders of information via social networks can assist in the acceleration or hindrance of information dissemination, in increased product exposure, and in the detection of contagious disease outbreaks. Hub nodes, high betweenness nodes, high closeness nodes, and high k-shell nodes have been identified as good initial spreaders. However, researchers have overlooked node diversity within network structures as a means of measuring spreading ability. The two-step framework described in this paper uses a robust and insensitive measure that combines global diversity and local features (e.g., degree centrality) to identify the most influential social network nodes. Preliminary experiment results indicate that the proposed method performs well and maintains stability in single initial spreader scenarios associated with different social network datasets.

Developing computer-controlled groups to engage in combat, control the use of limited resources, and create units and buildings in real-time strategy (RTS) games is a novel application in game AI. However, tightly controlled online commercial game pose challenges to researchers interested in observing player activities, constructing player strategy models, and developing practical AI technology in them. Instead of setting up new programming environments or building a large amount of agentpsilas decision rules by playerpsilas experience for conducting real-time AI research, the authors use replays of the commercial RTS game StarCraft to evaluate human player behaviors and to construct an intelligent system to learn human-like decisions and behaviors. A case-based reasoning approach was applied for the purpose of training our system to learn and predict player strategies. Our analysis indicates that the proposed system is capable of learning and predicting individual player strategies, and that players provide evidence of their personal characteristics through their building construction order.

A major task for postgenomic systems biology researchers is to systematically catalogue molecules and their interactions within living cells. Advancements in complex-network theory are being made toward uncovering organizing principles that govern cell formation and evolution, but we lack understanding of how molecules and their interactions determine how complex systems function. Molecular bridge motifs include isolated motifs that neither interact nor overlap with others, whereas brick motifs act as network foundations that play a central role in defining global topological organization. To emphasize their structural organizing and evolutionary characteristics, we define bridge motifs as consisting of weak links only and brick motifs as consisting of strong links only, then propose a method for performing two tasks simultaneously, which are as follows: 1) detecting global statistical features and local connection structures in biological networks and 2) locating functionally and statistically significant network motifs. To further understand the role of biological networks in system contexts, we examine functional and topological differences between bridge and brick motifs for predicting biological network behaviors and functions. After observing brick motif similarities between E. coli and S. cerevisiae, we note that bridge motifs differentiate C. elegans from Drosophila and sea urchin in three types of networks. Similarities (differences) in bridge and brick motifs imply similar (different) key circuit elements in the three organisms. We suggest that motif-content analyses can provide researchers with global and local data for real biological networks and assist in the search for either isolated or functionally and topologically overlapping motifs when investigating and comparing biological system functions and behaviors.

Robot intelligence architecture has advanced from action intelligence to autonomous intelligence, whereby robots can adapt to complex environments and interact with humans. This technology, considered central to next generation robots (NGRs), will become increasingly visible in many human service scenarios in the next two decades. Accordingly, there is an emerging need to predict and address intertwined technological and legal issues that will arise once NGRs become more commonplace. Safety issues will be of particular interest from a legal viewpoint. As robots become more capable of autonomous behavior, regulations associated with industrial robots will no longer be effective. In this paper we will discuss issues associated with autonomous robot behavior regulations associated with the concept of safety intelligence (SI). We believe the SI concept (one of several robot sociability problems) is crucial to the development of "robot law" that will accompany the establishment of a society in which humans and robots co-exist.

The authors describe a recommender model that uses intermediate agents to evaluate a large body of subjective data according to a set of rules and make recommendations to users. After scoring recommended items, agents adapt their own selection rules via interactive evolutionary computing to fit user tastes, even when user preferences undergo a rapid change. The model can be applied to such tasks as critiquing large numbers of music, image, or written compositions. In this paper we use musical selections to illustrate how agents make recommendations and report the results of several experiments designed to test the model’s ability to adapt to rapidly changing conditions yet still make appropriate decisions and recommendations.

Reinforcement learning theory encourages the use of agents for stimulating and assisting learners in their efforts to develop thinking styles. In this study the authors looked at a similar scenario of human-environmental interaction using Internet-mediated simulations as learning environments. One hundred and forty-nine vocational high schools students participated in this study to see if they can develop adequate thinking style when they learned in a simulation environment with help from agents. The results show that the judicial thinking style was best suited to the system we designed for this project - that is, we observed the greatest amount of development for this particular thinking style. Our results indicate that it is possible to establish and support thinking styles via Internet-mediated simulations.

Successful computer and Internet networks require carefully designed routing protocols. The authors report on their attempt to apply evolutionary computations - that is, to place a learning classifier system on individual routers - to solve routing problems. We found that learning classifier systems are capable of fulfilling traditional routing protocol tasks (e.g., establishing routing tables) after a short period of training. Furthermore, they are capable of adapting to changing network environments and choosing the most efficient path available. Results from our experiments show that the system outperforms shortest path algorithms.

Watts and Strogatz's "small world model" of disordered networks is becoming an increasingly popular research tool for modeling human society. As part of this approach, local information mechanisms (landscape properties) are used to approximate real-world conditions in social simulations. The authors investigate the influence of local information on social simulations that are performed using the small world model. In addition to defining local information, we use a cellular automata variation with added shortcuts as a test platform for simulating the spread of an epidemic and examining various influences. We believe our results help future researchers determine appropriate simulation parameters.

Azuaje offers an approach to the co-evolution of competing virtual creatures and a model for the evolution of game strategies and their emerging behaviors (F. Azuaje, see ibid., vol. 33, p.498-502, 2003). This model can be greatly simplified and optimal solutions can be obtained more quickly and easily by using an analytical approach. We emphasize the importance of performing a model analysis before choosing an evolutionary or analytical approach to a problem. Furthermore, Azuaje's model is derived from the Prisoner's Dilemma, a classical model in game theory; some results have already been discussed in the literature. We discuss his model from the perspective of game theorists.

Hysteresis is a memory effect frequently observed in physical research. The output of a hysteresis system is independent of input speed. This property, known as rate independence, significantly distinguishes hysteresis from short-term memory effects. This work conducts a numerical simulation to demonstrate that conventional models with short-term memories cannot properly simulate hysteresis trajectories. Subsequently, a novel model is developed to contribute to the field of system modeling. Experimental results confirm that the proposed model can model hysteresis behavior precisely.

Genetic algorithms (GAs) are a highly effective and efficient means of solving optimization problems. Gene encoding, fitness landscape and genetic operations are vital to successfully developing a GA. F. Cheong and R. Lai (see ibid., vol. 30, p. 31-46 (2000)) described a novel method, which employed an enhanced genetic algorithm with multiple populations, to optimize a fuzzy controller, and the experimental results revealed that their method was effective in producing a well-formed fuzzy rule-base. However, their encoding method and fitness function appear unnatural and inefficient. This study proposes an alternative method of concise genetic encoding and fitness design.

Hysteresis is a unique type of dynamic, which contains an important property, rate-independent memory. In addition to other memory-related studies such as time delay neural networks, recurrent networks, and reinforcement learning, rate-independent memory deserves further attention owing to its potential applications. In this paper, we attempt to define hysteretic memory (rate independent memory) and examine whether or not it could be modeled in neural networks. Our analysis results demonstrate that other memory-related mechanisms are not hysteresis systems. A novel neural cell, referred to herein as the propulsive neural unit, is then proposed. The proposed cell is based on a notion related the submemory pool, which accumulates the stimulus and ultimately assists neural networks to achieve model hysteresis. In addition to training by backpropagation, a combination of such cells can simulate given hysteresis trajectories.

Fundamental and advanced developments in neuro-fuzzy synergisms for modeling and control are reviewed. The essential part of neuro-fuzzy synergisms comes from a common framework called adaptive networks, which unifies both neural networks and fuzzy models. The fuzzy models under the framework of adaptive networks is called adaptive-network-based fuzzy inference system (ANFIS), which possess certain advantages over neural networks. We introduce the design methods for ANFIS in both modeling and control applications. Current problems and future directions for neuro-fuzzy approaches are also addressed.>

We summarize Jang's architecture of employing an adaptive network and the Kalman filtering algorithm to identify the system parameters. Given a surface structure, the adaptively adjusted inference system performs well on a number of interpolation problems. We generalize Jang's basic model so that it can be used to solve classification problems by employing parameterized t-norms. We also enhance the model to include weights of importance so that feature selection becomes a component of the modeling scheme. Next, we discuss two ways of identifying system structures based on Jang's architecture: the top-down approach, and the bottom-up approach. We introduce a data structure, called a fuzzy binary boxtree, to organize rules so that the rule base can be matched against input signals with logarithmic efficiency. To preserve the advantage of parallel processing assumed in fuzzy rule-based inference systems, we give a parallel algorithm for pattern matching with a linear speedup. Moreover, as we consider the communication and storage cost of an interpolation model. We propose a rule combination mechanism to build a simplified version of the original rule base according to a given focus set. This scheme can be used in various situations of pattern representation or data compression, such as in image coding or in hierarchical pattern recognition.>

It is shown that, under some minor restrictions, the functional behavior of radial basis function networks (RBFNs) and that of fuzzy inference systems are actually equivalent. This functional equivalence makes it possible to apply what has been discovered (learning rule, representational power, etc.) for one of the models to the other, and vice versa. It is of interest to observe that two models stemming from different origins turn out to be functionally equivalent.