Narrow Results

To achieve the exaFLOPS performance within a contained power budget, next generation supercomputers will feature hundreds of millions of components operating at low- and near-threshold voltage. As the probability that at least one of these components fails during the execution of an application approaches certainty, it seems unrealistic to expect that any run of a scientific application will not experience some performance faults. We believe that there is need of a new generation of light-weight performance and debugging tools that can be used online even during production runs of parallel applications and that can identify performance anomalies during the application execution. In this work we propose the design and implementation of a monitoring system that continuously inspects the evolution of running applications and the health of the system. To achieve minimum runtime overhead while maintaining the desired level of flexibility, we propose a decoupled approach in which accurate monitoring is performed at kernel-level while performance anomaly disambiguation and corrective actions are performed at user-level.

We evaluate our monitoring system on a 128-core AMD Interlagos system: First, we show that the runtime overhead of the monitoring system is negligible (0-2%). Then we show how our system can be used to precisely identify performance faults in three different scenarios: OS noise, application co-scheduling and dynamic power capping.

The influence of pump direction in all-optical gain-clamped fiber Raman amplifier using the directional couplers in feedback loop is demonstrated. The isolator direction decided the lasing direction in feedback loop. The comparison of performance in two lasing directions for both schemes is shown. It is concluded that the forward pumping configuration with forward lasing has better gain clamping performance and noise performance.

This paper investigates the pattern formation in a reaction–diffusion (R-D) system where two interacting species form coupled positive and negative feedback loops. It is found that the cooperation of competition and cross-diffusion can lead to the Turing pattern formation for which an adequate set of conditions are analytically derived. Such a mechanism of generating Turing patterns is different from the case that self-diffusion is sufficient to generate Turing patterns in a paradigm model (proverbially called as the Turing model) where two interacting species constitute a single negative feedback loop. Therefore, this work not only provides another model paradigm for studying the pattern formation but also would be helpful for understanding the formation of, for example, diversiform skin patterns in the mammalian world where coupled positive and negative feedback loops are ubiquitous.

By fairly simple considerations of stability and multistationarity in nonlinear systems of first order differential equations it is shown that under quite mild restrictions a negative feedback loop is a necessary condition for stability, and that a positive feedback loop is a necessary condition for multistationarity.

We state precisely and demonstrate two conjectures of R. Thomas following which (a) the existence of a positive circuit in the oriented interaction graph of a differential system is a necessary condition for the existence of several steady states, and (b) the existence of a negative non-oriented circuit of length at least two is a necessary condition for the existence of a stable periodic orbit.

The loss of cell cycle control is often associated with cancers and other different diseases. With the accumulation of omics data, the network for molecule interactions in the cell cycle process will become much clearer. The identification of the crucial modules in a giant network and investigation of inherent control relations are very important to the understanding of the molecular mechanisms of diseases for new drug design. The paper proposes novel techniques in analyzing such core regulatory modules based on network and system control theories. We initially define the degree of participation (DOP) and the rate of activity (ROA) for indentifying core module components, and then the diverse contribution elasticity functions for quantifying pairwise regulatory or control activities between those components, thus facilitating the decomposition of expanded core modules and the formation of feedback loops within the control schema. Motivated by the inherent regulatory mechanisms, we expound a kind of multiphase nonlinear adaptive control algorithm in repelling abnormal genetic mutations, which directly and indirectly impact cancer development in biological cells and organs. Experimental predictions are also elucidated within the work, helping those in vivo design, verification and performance evaluation.

The increasing number of knowledge-based systems that build on Bayesian networks and dynamic Bayesian networks acknowledges the usefulness of these frameworks for addressing complex real-life problems. The usually large number of probabilities required for their application, however, is often considered a major obstacle. The use of qualitative abstractions may to some extent remove this obstacle. Qualitative Bayesian networks and associated algorithms have been developed before. Based on qualitative Bayesian networks, in this paper, we present three definitions for the time-series dynamic Bayesian networks with feedback loops and qualitative time-series dynamic Bayesian networks by defining qualitative influence between adjacent time slices analogously. Then, we apply the qualitative dynamic Bayesian networks with feedback loops to an economic example and make qualitative decision successfully.

A gain-clamped fiber Raman amplifier is proposed and builds up. The result shows that gain fluctuation will be champed to a lower level with the dynamic gain control. And the dynamic gain control method is also fit for the Raman amplifiers with multi-wavelength pump.

The inherent micro-structure (agent-agent/system) of human organizations has been introduced in Chapters 2 and 3. Fundamentally, human organizations are composite complex adaptive systems with human beings as interacting agents (each an intrinsic complex adaptive system). This chapter further analyzes the basic conceptual foundation of the multi-layer structure, including advantages of the intelligent biotic macro-structure (with inherent features similar to that of an intelligent biological being — a structural reform), and its unique and more integrative complex adaptive dynamic in intelligent human organizations (towards iCAD). The necessity of nurturing an intelligent biotic macro-structure with vital characteristics that better synchronize and enhance sophisticated information/knowledge-related activities is highly beneficial — achieving a higher structural capacity. Thus, the attributes, functions and higher structural capacity of the more intelligent biotic macro-structure can reinforced the competitiveness of any categories of human organizations extensively.

In this respect, connecting and engaging of intelligence/consciousness sources (individual and collective), organizing around intelligence, intelligence/ consciousness management, and the intelligent biotic macrostructure are mutually enhancing (towards higher coherency). Apparently, being intelligence/consciousness-centric is a beneficial and critical activator (strategic foundation) of the intelligence paradigmatic shift. In the present context, the roles and integration of intelligence, information and knowledge, as well as nurturing a ‘common’ language and elevating coherency in human organizations (with respect to the macro-structure and micro-structure, as well as their higher collectiveness — collectiveness capacity) must be more deeply scrutinized and utilized. The presence and significance of the individual intelligence enhancer encompassing three entities namely, intelligence, knowledge, and theory in the human thinking systems, and the necessity of nurturing a similar and effective intelligence enhancer at organizational level are analyzed. Subsequently, the supporting roles and contributions of artificial intelligence systems are also examined.

In between the macro-structure and micro-structure are two meso-structures. In the intelligent organization theory, the complexity meso-structure encompasses spaces of complexity and punctuation points. In this respect, complexity is a highly significant focal point, and the exploitation of co-existence of order and complexity is a new necessity (complexity-centricity). Next, the network meso-structure encompassing complex network (network of networks) is also an inherent structure and dynamic in all human organizations. This meso-structure is briefly introduced, and will be more deeply analyzed with respect to governance (network-centricity, network governance).

Hence, it is crucial to lead and manage human organizations with a strategic approach that integrates the above multi-layer structure/ dynamic at all time so that a higher structural capacity, collectiveness capacity, adaptive capacity, self-organizing capacity, and emergence-intelligence capacity can be nurtured. In the current highly competitive context, possessing these positive capabilities to elevate coherency and synergetic characteristics (including social consensus and the construal aspect) and dynamic is also highly crucial — a key focus of the complexity-intelligence strategy (towards achieving higher organizational mental cohesion). Hence, the significance and impact of nurturing intelligent human organizations with the complexity-intelligence-centric and network-centric approach that leads to the emergent of smarter evolvers and emergent strategists must be better understood and adopted. (The conceptual foundation on structural-dynamic coherency and synergy in intelligent human organizations developed in this chapter will be more deeply reviewed and exploited in later chapters.)