Reservoir Network Research Articles

Heterogeneity quantification in waterfloods using a multiphase network approach

Knowledge of reservoir heterogeneity in waterflooding is key to achieving a better sweep efficiency in a waterflooding process. However, a high degree of spatial variation in reservoir properties and a limited number of measurements make heterogeneity quantification difficult. Furthermore, the contributions of heterogeneities to waterflooding efficiency are unknown. We present a multiphase reservoir network model that can not only approximate interwell heterogeneity, but that can also be used to improve the performance of the waterflood.We develop the network model by dividing the reservoir into a number of node volumes, each of which contains a well. Flow between the node volumes is obtained by solving for the pressure field in the diffusivity equation. A network of pseudo-streamlines connecting wells (producer-injector and injector-injector) is then generated in order to map the saturation between the wells. Instead of grid properties, flow area and interwell Dykstra-Parsons coefficients for time of flight characterize flow and saturation along the streamlines. The model parameters are estimated by a history matching of the fractional-flow data. A derivative-free optimization algorithm is then used to minimize the objective function (the sum of the squared differences between the simulated and actual producer fractional flow).The approach that we present uses only injection and production rates, well positions, and well fractional-flow values, which are the most accessible data in a field. The method was tested using several reservoir models with different synthetic permeability distributions and geologic features. Results show that as interwell heterogeneity increases, Dykstra-Parsons coefficients increase, and flow areas decrease. Furthermore, frontal shock, which is a characteristic of displacement of immiscible incompressible phases, is prevented by the introduction of Dykstra-Parsons coefficients.Flow area and the time-of-flight Dykstra-Parson coefficient can be used to improve waterflood sweep efficiency. Resulting parameters also help in decision making regarding injected water reallocation, but only through a study of injection and production rates.

PLXTRM: Prediction-Led eXtended-guitar Tool for Real-time Music applications and live performance

This article presents PLXTRM, a system tracking picking-hand micro-gestures for real-time music applications and live performance. PLXTRM taps into the existing gesture vocabulary of the guitar player. On the first level, PLXTRM provides a continuous controller that doesn’t require the musician to learn and integrate extrinsic gestures, avoiding additional cognitive load. Beyond the possible musical applications using this continuous control, the second aim is to harness PLXTRM’s predictive power. Using a reservoir network, string onsets are predicted within a certain time frame, based on the spatial trajectory of the guitar pick. In this time frame, manipulations to the audio signal can be introduced, prior to the string actually sounding, ’prefacing’ note onsets. Thirdly, PLXTRM facilitates the distinction of playing features such as up-strokes vs. down-strokes, string selections and the continuous velocity of gestures, and thereby explores new expressive possibilities.

Coupling ecological and social network models to assess “transmission” and “contagion” of an aquatic invasive species

Network analysis is used to address diverse ecological, social, economic, and epidemiological questions, but few efforts have been made to combine these field-specific analyses into interdisciplinary approaches that effectively address how complex systems are interdependent and connected to one another. Identifying and understanding these cross-boundary connections improves natural resource management and promotes proactive, rather than reactive, decisions. This research had two main objectives; first, adapt the framework and approach of infectious disease network modeling so that it may be applied to the socio-ecological problem of spreading aquatic invasive species, and second, use this new coupled model to simulate the spread of the invasive Chinese mystery snail (Bellamya chinensis) in a reservoir network in Southeastern Nebraska, USA. The coupled model integrates an existing social network model of how anglers move on the landscape with new reservoir-specific ecological network models. This approach allowed us to identify 1) how angler movement among reservoirs aids in the spread of B. chinensis, 2) how B. chinensis alters energy flows within individual-reservoir food webs, and 3) a new method for assessing the spread of any number of non-native or invasive species within complex, social-ecological systems.

A large scale network model to obtain interwell formation characteristics

Limited data availability and poor data quality make it difficult to characterise many reservoirs. For waterflooded reservoirs, production and injection data provide information from which injector-to-producer connections can be inferred. In this research, well locations and injection and production rate data are used to develop a reservoir-scale network model. A Voronoi mesh divides the reservoir into node volumes, each of which contains a well. Bonds connect the nodes with conductance values that are inferred from the rate data. The inverse problem minimises the mean-squared difference between computed and observed production data by adjusting the conductances between nodes. A derivative free optimisation algorithm is used to minimise the mean-squared difference. This coarse network model approach is fast and efficient because it solves for a small number of unknowns and is less underdetermined than correlation-based methods. The reservoir network model has promise as a reservoir description tool because of its modest data requirements, flexibility, efficiency, interpretability, and dynamism. [Received: July 17, 2015; Accepted: January 14, 2016]

A large scale network model to obtain interwell formation characteristics

Limited data availability and poor data quality make it difficult to characterise many reservoirs. For waterflooded reservoirs, production and injection data provide information from which injector-to-producer connections can be inferred. In this research, well locations and injection and production rate data are used to develop a reservoir-scale network model. A Voronoi mesh divides the reservoir into node volumes, each of which contains a well. Bonds connect the nodes with conductance values that are inferred from the rate data. The inverse problem minimises the mean-squared difference between computed and observed production data by adjusting the conductances between nodes. A derivative free optimisation algorithm is used to minimise the mean-squared difference. This coarse network model approach is fast and efficient because it solves for a small number of unknowns and is less underdetermined than correlation-based methods. The reservoir network model has promise as a reservoir description tool because of its modest data requirements, flexibility, efficiency, interpretability, and dynamism. [Received: July 17, 2015; Accepted: January 14, 2016]

A soft, wearable microfluidic device for the capture, storage, and colorimetric sensing of sweat.

Capabilities in health monitoring enabled by capture and quantitative chemical analysis of sweat could complement, or potentially obviate the need for, approaches based on sporadic assessment of blood samples. Established sweat monitoring technologies use simple fabric swatches and are limited to basic analysis in controlled laboratory or hospital settings. We present a collection of materials and device designs for soft, flexible, and stretchable microfluidic systems, including embodiments that integrate wireless communication electronics, which can intimately and robustly bond to the surface of the skin without chemical and mechanical irritation. This integration defines access points for a small set of sweat glands such that perspiration spontaneously initiates routing of sweat through a microfluidic network and set of reservoirs. Embedded chemical analyses respond in colorimetric fashion to markers such as chloride and hydronium ions, glucose, and lactate. Wireless interfaces to digital image capture hardware serve as a means for quantitation. Human studies demonstrated the functionality of this microfluidic device during fitness cycling in a controlled environment and during long-distance bicycle racing in arid, outdoor conditions. The results include quantitative values for sweat rate, total sweat loss, pH, and concentration of chloride and lactate.

Driving reservoir models with oscillations: a solution to the extreme structural sensitivity of chaotic networks.

A large body of experimental and theoretical work on neural coding suggests that the information stored in brain circuits is represented by time-varying patterns of neural activity. Reservoir computing, where the activity of a recurrently connected pool of neurons is read by one or more units that provide an output response, successfully exploits this type of neural activity. However, the question of system robustness to small structural perturbations, such as failing neurons and synapses, has been largely overlooked. This contrasts with well-studied dynamical perturbations that lead to divergent network activity in the presence of chaos, as is the case for many reservoir networks. Here, we distinguish between two types of structural network perturbations, namely local (e.g., individual synaptic or neuronal failure) and global (e.g., network-wide fluctuations). Surprisingly, we show that while global perturbations have a limited impact on the ability of reservoir models to perform various tasks, local perturbations can produce drastic effects. To address this limitation, we introduce a new architecture where the reservoir is driven by a layer of oscillators that generate stable and repeatable trajectories. This model outperforms previous implementations while being resistant to relatively large local and global perturbations. This finding has implications for the design of reservoir models that capture the capacity of brain circuits to perform cognitively and behaviorally relevant tasks while remaining robust to various forms of perturbations. Further, our work proposes a novel role for neuronal oscillations found in cortical circuits, where they may serve as a collection of inputs from which a network can robustly generate complex dynamics and implement rich computations.

Synoptic climatology of the early 21st century drought in the Colorado River Basin and relationships to reservoir water levels

ABSTRACTSince the year 2000, streamflows throughout the Colorado River Basin (CRB) in the southwestern United States have been greatly impacted by a prolonged drought. As a primarily mountain snowmelt‐sourced basin, the climatic drivers of annual runoff in the CRB bear significant influence on the hydrologic status of its network of reservoirs. To analyse reservoir response to the drought, this study takes a holistic approach of relating the frequency of mid‐tropospheric circulation patterns (CPs) to water levels at Lake Powell and Lake Mead, the primary reservoirs in the CRB. A synoptic climatology of the region was constructed by classifying daily mean 500‐hPa geopotential height values from 1982 to 2012, a span that includes the early 21st century drought and the relatively wetter years preceding it, to determine how the drought manifests itself in the trends of CPs over time. The CPs were then statistically compared to several measures of reservoir water level, including yearly inflow and net changes in storage year‐to‐year, to examine potential relationships. The results of the classification reveal that the CRB is most notably characterized by dominant summertime ridge patterns, which have become slightly more common during the drought. Cool season troughs were generally linked with the presence of more water in the reservoirs, but ridges have stronger relationships to reservoir variables, including linkages to years with lower reservoir inflow, greater yearly losses, and net declines in storage year‐to‐year. Further, dry transitional patterns, CPs which represent zonal flow yielding little precipitation, are increasingly common in the spring, coming at the expense of wetter CPs. These results support previous research regarding the observed hastening of snowmelt, contributing to the early 21st century drought in the CRB as well as the overall trend in declining mountain snowpack across western North America.

Bidirectional reservoir networks trained using SVM $$+$$ + privileged information for manufacturing process modeling

In the last decade, a wide range of machine learning approaches were proposed and experimented to model highly nonlinear manufacturing processes. However, improving the performance of such models is challenging due to the complexity and high dimensionality of the manufacturing processes in general. In this paper, we propose bidirectional echo state reservoir networks (Bi-ESNs) trained using support vector machine privileged information method (SVM\(+\)) to model a winding machine process. The proposed model will be applied, tested and compared to reported models in the literature such as classical ESN with linear regression, ESN with a linear SVM readout, genetic programming, feedfoward neural network with backpropagation, radial basis function network, adaptive neural fuzzy inference system and local linear wavelet neural network. The developed results show that Bi-ESNs trained with SVM\(+\) are promising. It was able to provide better generalization performance compared to other models.

Reservoir Computing Properties of Neural Dynamics in Prefrontal Cortex.

Primates display a remarkable ability to adapt to novel situations. Determining what is most pertinent in these situations is not always possible based only on the current sensory inputs, and often also depends on recent inputs and behavioral outputs that contribute to internal states. Thus, one can ask how cortical dynamics generate representations of these complex situations. It has been observed that mixed selectivity in cortical neurons contributes to represent diverse situations defined by a combination of the current stimuli, and that mixed selectivity is readily obtained in randomly connected recurrent networks. In this context, these reservoir networks reproduce the highly recurrent nature of local cortical connectivity. Recombining present and past inputs, random recurrent networks from the reservoir computing framework generate mixed selectivity which provides pre-coded representations of an essentially universal set of contexts. These representations can then be selectively amplified through learning to solve the task at hand. We thus explored their representational power and dynamical properties after training a reservoir to perform a complex cognitive task initially developed for monkeys. The reservoir model inherently displayed a dynamic form of mixed selectivity, key to the representation of the behavioral context over time. The pre-coded representation of context was amplified by training a feedback neuron to explicitly represent this context, thereby reproducing the effect of learning and allowing the model to perform more robustly. This second version of the model demonstrates how a hybrid dynamical regime combining spatio-temporal processing of reservoirs, and input driven attracting dynamics generated by the feedback neuron, can be used to solve a complex cognitive task. We compared reservoir activity to neural activity of dorsal anterior cingulate cortex of monkeys which revealed similar network dynamics. We argue that reservoir computing is a pertinent framework to model local cortical dynamics and their contribution to higher cognitive function.

Site‐Selective In Situ Grown Calcium Carbonate Micromodels with Tunable Geometry, Porosity, and Wettability

Micromodels with simplified porous microfluidic systems are widely used to mimic the underground oil‐reservoir environment for multiphase flow studies, enhanced oil recovery, and reservoir network mapping. However, previous micromodels cannot replicate the length scales and geochemistry of carbonate because of their material limitations. Here a simple method is introduced to create calcium carbonate (CaCO3) micromodels composed of in situ grown CaCO3. CaCO3 nanoparticles/polymer composite microstructures are built in microfluidic channels by photopatterning, and CaCO3 nanoparticles are selectively grown in situ from these microstructures by supplying Ca2+, CO32− ions rich, supersaturated solutions. This approach enables us to fabricate synthetic CaCO3 reservoir micromodels having dynamically tunable geometries with submicrometer pore‐length scales and controlled wettability. Using this new method, acid fracturing and an immiscible fluid displacement process are demonstrated used in real oil field applications to visualize pore‐scale fluid–carbonate interactions in real time.

Reservoir impacts downstream in highly regulated river basins: the Ebro delta and the Guadalquivir estuary in Spain

Abstract. Regulation by reservoirs affects both the freshwater regime and the sediment delivery at the area downstream, and may have a significant impact on water quality in the final transitional water bodies. Spain is one the countries with more water storage capacity by reservoirs in the world. Dense reservoir networks can be found in most of the hydrographic basins, especially in the central and southern regions. The spatial redistribution of the seasonal and annual water storage in reservoirs for irrigation and urban supply, mainly, has resulted in significant changes of water flow and sediment load regimes, together with a fostered development of soil and water uses, with environmental impacts downstream and higher vulnerability of these areas to the sea level rise and drought occurrence. This work shows these effects in the Guadalquivir and the Ebro River basins, two of the largest regulated areas in Spain. The results show a 71 % decrease of the annual freshwater input to the Guadalquivir River estuary during 1930–2014, an increase of 420 % of the irrigated area upstream the estuary, and suspended sediment loads up to 1000 % the initial levels. In the Ebro River delta, the annual water yield has decreased over a 30 % but, on the contrary, the big reservoirs are located in the main stream, and the sediment load has decreased a 99 %, resulting in a delta coastal regression up to 10 m per year and the massive presence of macrophytes in the lower river. Adaptive actions proposed to face these impacts in a sea level rise scenario are also analyzed.

Estimation of water saturation by using radial based function artificial neural network in carbonate reservoir: A case study in Sarvak formation

Water saturation determination in core laboratory is known as a cost and time consuming labor. Hitherto, many scientists attempted to estimate accurately water saturation from well-logging data which has a continuous record without losing information. Therefore, various model were introduced to relate reservoir properties and water saturation. Since carbonate reservoir is very heterogeneous in shape and size of pore throat, the relation between water saturation and other carbonates reservoir properties is very complex, and causes considerable overall errors in water saturation calculation. By increasing the usage and improvement of soft computing methods in engineering problems, petroleum engineers have been attended them to measure the petrophysical properties of the reservoir. In this study, a radial basis function neural network (RBFNN) improved by genetic algorithm has been employed to estimate formation water saturation by using conventional well-logging data. The used logging and core data have been gathered from a carbonated formation from one of oilfield located in south-west Iran, and finally their results of the proposed model were compared with the core analysis results. By checking the testing data from another well, it showed this method had a 0.027 for mean square errors and its correlation coefficient is equal to 0.870. These results implied on high accuracy of this model for oil saturation degree estimation. While the common methods like Archie, had a 0.041 mean square error and 0.720 of the correlation coefficient, which indicate a high ability of RBF model than the other usual empirical methods.

The Influence Between Natural Fracture and Hydraulic Fracture in Shale Reservoir

Based on linear elastic fracture mechanics, the forming mechanism of complex fracture system in shale reservoir are studied. The interference of the natural fracture and hydraulic fracture model and a new crack initiation angle calculation model is established, and the influence factors of critical differential stress level and new crack initiation angle are analyzed. The impact of natural fracture dip and natural fracture density on complex fracture network are researched. Calculation results show that the lower approximation Angle and high pressure in the fracture is the necessary conditions of forming fracture network; Under the condition of lower stress difference and high internal pressure, the new crack can be faster to the maximum horizontal in-situ stress direction, connect more natural fracture; in order to get a complex fracture network in shale reservoir in the need to constantly improve the pressure to activate and connect higher approaching Angle and distance of natural fracture. The research results have important reference value in understanding the formation mechanism of complex fracture network in shale reservoir, can also be used for shale fracturing design and fracturing interval optimization.

Numerical simulation of CO2 huff-n-puff in complex fracture networks of unconventional liquid reservoirs

Horizontal drilling and hydraulic fracturing produce Unconventional Liquid Reservoirs (ULRs) economically. One of the characteristics of the ULRs is high production declining rate, and thus enhanced oil recovery techniques such as Huff-n-Puff will ensue. In addition, the existence of complex fracture network has been proven by both field and simulation work. This study investigated the potential of CO2 huff-n-puff in ULRs with complex fracture networks. Two fracture characterization techniques for complex fracture network such as fractal-based, and microseismic-based were applied to reduce uncertainties of fracture networks and thus improve the accuracy of simulation input. In order to study complex CO2 recovery mechanism, lab-measured data was upscaled to the field-scale, and then used for evaluating production performance of a simple fracture network. Detailed sensitivity analysis showed that diffusion in the field scale is negligible compared with other recovery mechanisms such as oil swelling, gas expansion, viscosity reduction, etc. The following comparison between dual continuum and the explicit fracture modeling approaches highlights the importance of heterogeneity due to the explicit fractures. Dual continuum models yields less accurate pressure behavior than explicit discrete fracture models even though production behavior could be history matched.

Microfluidics-Based PCR for Fusion Transcript Detection.

The microfluidic technology allows the production of network of submillimeter-size fluidic channels and reservoirs in a variety of material systems. The microfluidic-based polymerase chain reaction (PCR) allows automated multiplexing of multiple samples and multiple assays simultaneously within a network of microfluidic channels and chambers that are co-ordinated in controlled fashion by the valves. The individual PCR reaction is performed in nanoliter volume, which allows testing on samples with limited DNA and RNA. The microfluidics devices are used in various types of PCR such as digital PCR and single molecular emulsion PCR for genotyping, gene expression, and miRNA expression. In this chapter, the use of a microfluidics-based PCR for simultaneous screening of 14 known fusion transcripts in patients with leukemia is described.

Optimal Operation of a Network of Multi-purpose Reservoir: A Review

Due to the effects of climate change and population growth, reservoirs play a more and more important role in water resources management. The management of a multi-reservoir system is complex due to the curse of dimensionalities, nonlinearities and conflicts between different objectives. The optimal operation of a multi-reservoir system operation typically involves optimization and simulation models, which can provide the quantitative information to improve operational water management. The objectives of this paper are to extend previous state-of-the-art reviews in the operational management of a network of multi-purpose reservoirs with recent developments and to focus on the application of Model Predictive Control for real time control of a reservoir system.

A reservoir network model for sensory-guided probabilistic decision making

Animals are often required to adequately respond to novel stimuli on the basis of their previous sensory experiences. To investigate the neural mechanism underlying this sensory-guided decision making, we conducted multiunit recordings from the rats performing a two-alternative choice task. The rats were trained to make a LEFT or a RIGHT choice in response to two auditory cues, and then their behavior was tested for these familiar cues and other novel cues. Their choice probabilities generally varied in a graded manner such that the probability of choosing an option changed gradually according to frequency differences between familiar and novel cues. However, we also observed large variability in choice behavior across rats: choice probabilities for novel cues were near the chance level in some rats, while it systematically changed with tone frequency in other rats. We elucidated the mechanism underlying such decision behavior and the possible origin of individual behavioral differences in a neural network model (Fig. ​(Fig.1A).1A). Our model is viewed as a reservoir network [1] and learns to associate two familiar cues with two alternative choices through reinforcement learning with eligibility trace [2]. Figure 1 Our model successfully replicated the gradual choice behavior observed in our experiment. We further analyzed how neural dynamics in the reservoir network determines the choice probability for familiar and novel cues. We found that a familiar stimulus sequentially activates a relatively small portion of reservoir neurons, and reinforcement learning trained output connections from these neurons such that only an adequate output neuron is activated by the neural trajectory evoked in the reservoir. We further revealed that choice responses to novel cues become graded due to trial-by-trial overlaps between the familiar trajectories and novel-cue-evoked trajectories. Interestingly, our model also exhibited similar variability in choice responses to that observed in individual rats. We found that if input neurons are highly sensitive to external stimuli, that is, if the width of their frequency tuning curves is broad, the model network likely shows gradual choice behavior for novel stimuli. In contrast, the model with more sensitive input neurons (with broader tuning curves) tends to generate near-random choice behavior, displaying a flat dependence of choice probability on novel stimuli. We compared choice behavior between the models and the rats by introducing quantitative measures and found that the behavioral tendency of our model is consistent with that of the rats (Fig. ​(Fig.1B1B). These results may suggest that some individual differences in decision making behavior emerge from neural population dynamics rather than differences in higher-level behavioral strategies.

Towards a biological plausible model of the interaction of long-term memory and working memory

Learning and memorizing past information are critical features of neural networks. Several models have been proposed to explain the storage and retrieval of different kinds of information [1,2] . These models differ in terms of the types of memories they can store, in terms of the time scales on which they operate, and in terms of the persistence of the stored information. In particular, there are neural models for the long-term memory (LTM) system [2] and other neural models for the working memory (WM) system [1]. While there are some controversies whether LTM and WM are realized by distinct brain regions, there is no doubt about the fact that these two systems are at the same time tightly coupled and interact continuously [3]. However, it is still widely unclear how this interaction is actually realized. Here, we demonstrate a model reproducing and explaining the interaction between LTM and WM such as to exploit the abilities and advantages of the individual components. WM is most probably realized in the prefrontal cortex (PFC). To capture the diverse dynamics observed in PFC during WM experiments, we choose to model WM by a reservoir computing network [4]. Basically, it consists of a large reservoir of randomly and recurrently connected neurons coupled to the input signals. Due to the huge variety of different signals present in the reservoir, these networks can serve as universal function approximators. However, the influence of past inputs fades away within time scales of seconds to minutes. The hippocampus, especially its CA3 area, is believed to play a major role in LTM formation on time scales of hours to days. We model its auto-associative capabilities by the well-known idea of cell assemblies [5]. According to this idea, declarative knowledge translates into correlations of neural activities which, in turn, lead to a potentiation of excitatory synapses. This process forms groups of highly interconnected neurons - named cell assemblies. If the synapses in a recurrent network are governed by Hebbian plasticity, cell assemblies emerge in an unsupervised manner based on correlations observed in the network input. Experimental findings indicate that rather abstract or general LTM knowledge does not depend on the hippocampus but is transferred to the medial prefrontal cortex (mPFC). In our model, the representations in the mPFC are the central interface between LTM and WM. It receives signals from both the cell-assembly as well as the reservoir network. Analyzing the capabilities and characteristics of the proposed system, we train the reservoir to perform certain arithmetic operations within different contexts. During operation, the cell assembly network learns to detect and remember associations between context signals and computed results. This enables the system to remember results of earlier calculations beyond the WM capacity and to reuse them for later tasks. Thus, the system is able to perform computations based on short-term memories of the operands as well as to overcome the limits of the short-term memory by storing computed results into the long-term memory and transferring them back for further usage.

Self-organization of computation in neural systems by interaction between homeostatic and synaptic plasticity

The ability to perform complex motor control tasks is essentially enabled by the nervous system via the self-organization of large groups of neurons into coherent dynamic activity patterns. During learning, this is brought about by synaptic plasticity, resulting in the formation of multiple functional networks - commonly termed 'cell-assemblies'. A multitude of such cell assemblies provide the requisite machinery for non-linear computations needed for the mastery of a large number of motor skills. However, given the fact that there exists considerable overlap between the usage of the same neurons within such assemblies, for a wide range of motor tasks, creation and sustenance of such computationally powerful networks poses a challenging problem. How such interwoven assembly networks self-organize and how powerful assemblies can coexist therein, without catastrophically interfering with each other remains largely unknown. One the one side, it is already known that networks can be trained to perform complex nonlinear calculations [1], such that, if the network possesses a reservoir of rich, transient dynamics, desired outputs can be extracted from these reservoirs in order to enable motor control. On the other side, cell assemblies are created by Hebbian learning rules that strengthen a synapse if pre- and post-synaptic neurons are co-active within a small enough time window [2]. Therefore it appears relatively straightforward to combine these mechanisms in order to construct powerful assembly networks. However, given that the self-organization of neurons into cell assemblies by the processes of synaptic plasticity induces ordered or synchronized neuronal dynamics, which can destroy the required complexity of a reservoir network, such a combination remains a very challenging problem [3]. Furthermore, simultaneous creation of multiple cell assemblies can also lead to catastrophic interference if one cannot prevent them from growing into each other. In this study, we exploit for the first time the interaction between neuronal and synaptic processes acting on different time scales to enable, on a long time scale, the self-organized formation of assembly networks (Fig. ​(Fig.1),1), while on a shorter timescale, to conjointly perform several non-linear calculations needed for motor fine-control. Specifically, by the combination of synaptic plasticity and synaptic scaling [4], as a homeostatic mechanism, we demonstrate that such self-organization allows executing a difficult, six degrees of freedom, manipulation task with a robot where assemblies need to learn computing complex nonlinear transforms and - for execution - must cooperate with each other without interference. This mechanism, thus, permits for the first time, the guided self-organization of computationally powerful sub-structures in dynamic networks for behavior control. Furthermore, comparing our assembly network to networks with unchanging synapses (static networks) shows that it is indeed the embedding of a strongly connected assembly that creates the necessary computational power. Figure 1 Cell assembly size and computational performance are correlated. (A) Input-driven formation of cell assemblies brought about by the interaction long-term potentiation (LTP) and synaptic scaling (Syn. Sca.). (B) With more learning trials the assembly grows ...