### Table 1 Deterministic chaos versus stochastic chaos

"... In PAGE 8: ... Following earlier convention, we will use the terminology Stochas- tic Chaos (SC) for the description of aperiodic behavior in brain dynamics which has been mod- eled by KIII (Freeman, 2000b; Kozma and Free- man, 2001; Werbos, 2000). Table1 summarizes our present understanding on the relation between deterministic chaos and stochastic chaos. SR has three main components, a bi- or multi- stable energy function, weak periodic) input sig- nal, and a noise component (Gammaitoni et al.... ..."

### Table 2: Time variables and their distribution functions. Time sample post-processing is handled in two di erent ways, depending on whether the un- derlying variable has a deterministic component. We summarize the two methods here: Pure stochastic variable: In this case the time-sample list is sorted as a histogram { divided into either 103 or 104 buckets (depending on the range and variation of the recorded process). Then, the histogram is normalized to the interval [0; 1], which yields a (synthesized) discrete probability distribution function (or pdf) f(t) for the variable, where we now assume that a given outcome is made as a simple Bernoulli decision. I.e., f(t) returns the probability of a sample time t being realized during playback on the device. Next, f apos;s cumulative distribution function F(t) is produced, and the output of the entire process is F?1(u), the CDF apos;s inverse transform, where u is uniformly distributed in [0; 1]. The simulator uses this function to generate random response times, in concert with a dedicated random number generator. Deterministic/stochastic variable: The sample trace is linearized by its deterministic index, 8

1998

Cited by 2

### Table 1: Classi cation of clock synchronization algorithms aD stands for deterministic, P for probabilistic and S for statistical. bComponents are labeled C (Clock), L (Link) and P(Processor). R stands for Reliable, C for Crash, P for Performance, B for Byzantine, and T for Timing. For probabilistic and statistical algorithms, no upper bound on message delays is assumed. Consequently, omission and performance failures are irrelevant, which is indicated by sign . cSymmetric and asymmetric schemes are denoted by SYM and ASYM. FLOOD, RING, MAST and SLAV name ooding-based, ring-based, master-controlled and slave-controlled

"... In PAGE 20: ....3.1 Classi cation of clock synchronization algorithms This section presents the classi cation of the most referenced clock synchronization algo- rithms. Table1 gives for each algorithm its type, according to the synchrony model de ned in section 2.2; the failure mode assumed for each component, as de ned in section 2.... In PAGE 20: ... MSG) is used when rounds are detected thanks to initially synchronized clocks (resp. message exchanges) Table1 shows that all the clock synchronization algorithms, whatever their type, can be completely described with the three building blocks we have identi ed in section 4.2, and the techniques implementing them.... ..."

### Table 2: Variability of the Major Interrupt Handler apos;s Execution Time On the loaded component the in uence of message tra c on the execution time of the major interrupt handler can be demonstrated: A communication slot on the bus has the length of 8 ms. During each slot one component and its redundant copy (which was switched o during our experiments) can send their messages. The Rolling Ball application uses 5 components which share the bus using a deterministic time-slicing protocol with a period of 5 8 = 40 ms. This corresponds to 5 executions of the major interrupt handler. As can be seen in Figure 9 every fth invocation of the major interrupt handler takes exceptionally long. This is caused by the actions that have to be performed after the observed component sends a message.

1991

Cited by 5

### Table 7 shows larger instances where n1 = n2 = n3 = 10. RandGen was expected for 500 iterations and GraspGen was executed for three outer iterations due to the extreme time requirements of the local search component. Figure 7 shows that GRASP is again competitive for longer runs, but dominated for shorter CPU times. The table indicates that the deterministic heuristics are much faster, but not reliable.

2003

"... In PAGE 20: ...8 / 7110 (time to nd sol.) (2477) (5085) (2214) (2767) (7110) Table7 : General Costs Obj. / CPU (sec) for 10 10 10, j j=30, =20, B(1) = 2, B(2) = 4 0 0.... ..."

Cited by 2

### Table 1: Classi cation of clock synchronization algorithms aD stands for deterministic, P for probabilistic and S for statistical. bComponents are labelled C (Clock), L (Link) and P (Processor). R stands for Reliable, C for Crash, P for Performance, B for Byzantine, and T for Timing. For probabilistic and statistical algorithms, no upper bound on message delays is assumed. Consequently, omission and performance failures are irrelevant, which is indicated by sign |. cSymmetric and asymmetric schemes are denoted by SYM and ASYM. FLOOD, RING, MAST and SLAV name ooding- based, ring-based, master-controlled and slave-controlled schemes. dNAV is used for non averaging techniques. Averaging techniques are denoted by AV and the convergence function name is given. eSYNC (resp. MSG) is used when rounds are detected thanks to initially synchronized clocks (resp. message exchanges)

"... In PAGE 18: ...ection 5.2, their performance and cost. 5.1 Classi cation of clock synchronization algorithms Table1 presents the classi cation of the most referenced clock synchronization algorithms. This table shows that all the clock synchronization algorithms, whatever their type, can be... In PAGE 20: ... 5.3 Concluding remarks While deterministic, probabilistic and statistical clock synchronization algorithms use the same building blocks (see Table1 ), the main di erence between these three classes of algo- rithms concerns their respective objectives. Deterministic algorithms aim at guaranteeing a worst-case precision.... ..."

### Table 1. Deterministic Propagation

2006

Cited by 2

### Table 1 Comparison of mean square errors of ROMKF and RUKF

1997

"... In PAGE 3: ...Table 1 Comparison of mean square errors of ROMKF and RUKF (MSE) of the images ltered with di erent ( 2 w1) values are shown in Table1 . For di erent observation noise vari- ances, the `deterministic component apos; error variance com- puted with the proposed method converged to f8, 22, 50, 104g, and the resulting MSE is indicated by a `* apos;.... ..."

Cited by 2

### Table 1: Inference on unit roots for univariate series Null Variables

"... In PAGE 4: ...xn = + A1 xn?1 + . . . + Ap?1 xn?p+1 ? xn?p + ut: (1) In Table1 results obtained from HEGY{regressions and from an application of the Jo- hansen procedure to a vector model like (1) are reported. The HEGY regressions contained seasonal dummy variables as deterministic components.... ..."

### Table 2: Embedded deterministic patterns

2000

"... In PAGE 4: ...3 Deterministic test patterns The efficiency of the deterministic BIST scheme strongly depends on the number of specified bits in deter- ministic patterns. Table2 provides figures on the determi- nistic patterns for the industrial circuits: the number of embedded deterministic patterns (#det. pats), the number of pseudo-primary inputs (#PPI), and the maximum num- ber (#spec.... In PAGE 4: ...Table 2: Embedded deterministic patterns Table2 indicates that the number of specified bits is quite small and does not generally increase with the size of the circuit or the number of pseudo-primary inputs. A similar observation has been reported with the ISCAS apos;85 and ISCAS apos;89 benchmark circuits in [HRTW95].... ..."

Cited by 9