Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 1.— 31-GHz CBI contours overlaid on a composite 3-color image of RCW 49: IRAC 3.6 µm (blue) and 8 µm (green) from the GLIMPSE survey, PACS 160 µm (red) from the Herschel OT2 program ”Unveiling the misterious case of RCW 49: a powerful HII region with associated Anomalous Microwave Emission” (PI.: R. Paladini). The IRAC 3.6 µm band shows the stellar content of the region, including the bright WR20b and Westerlund 2 cluster. The 8 µm band is dominated by the 7.7 µm PAH feature, and it highlights the Photo Dissociation Region (PDR) associated with RCW 49. Finally, the PACS 160 µm data evidence the presence of cold dust, both in the PDR and in its surroundings. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | follow-up on the original CBI detection, in order to shed light on the nature of the reported microwave excess. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | RCW 49, located at (RA, Dec) = (10h 24m 14.6s, -57◦ 46(cid:48) 58”), is the most luminous and massive HII region of the Southern Galaxy (e.g. Paladini et al. 2003), with a bolometric luminosity of 1.4 × 107 L(cid:12) and a stellar mass of ∼ 3 × 104 M(cid:12) (Vacca et al. 1996). The source has an estimated age of 2-3 Myr (Piatti et al. 1998), while its distance remains uncer- tain, with different authors suggesting a distance anywhere between 2.3 (e.g., Brand & Blitz 1993) and 6 – 7 kpc (e.g., Benaglia et al. 2013). RCW 49 extends for ∼ 2 degrees and is characterized by a very complex morphology. At its center is the Westerlund 2 (Wd2, Westerlund 1960) compact cluster, comprising a dozen of OB stars. An additional set of about 30 new OB star candidates is found in and around Wd2. Beyond the cluster core there are three more massive stars, i.e. a star of spectral type O4 or O5 (Rauw et al. 2007), a binary Wolf- Rayet star, WR20a, which is the most massive binary sys- tem in the Galaxy with a well-determined mass (Rauw et al. 2005), and another Wolf-Rayet star, WR20b (van der Hucht 2001), situated several arcminutes away. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | IR wavelengths. In the radio continuum, data are available at 408 MHz (73 cm, Shaver & Goss 1970b), 2.7 GHz (11 cm, Day et al. 1972), 5 GHz (6 cm, Caswell & Haynes 1987), 8.7 GHz (3.4 cm, McGee et al. 1975) and 14.7 GHz (2 cm, McGee & Newton 1981). All these observations are charac- terized by angular resolutions ranging from 8.2’ (at 2.7 GHz) to ∼ 2’ (e.g 14.7 GHz). The Molonglo Observatory Synthesis Telescope (MOST) carried out sub-arcmin resolution obser- vations at 0.843 GHz (35.6 cm, Whiteoak et al. 1989), and more recently the ATCA targeted the core of RCW 49 at 1.4 and 2.4 GHz (21 cm and 12.5 cm, Whiteoak & Uchida 1997, hereafter WU97), as well as at 5.5 and 9.0 GHz (5.4 cm and 3.3 cm, Benaglia et al. 2013). These observations span an- gular resolutions from 10” to 2”. All these data show that the core of RCW 49 is dominated by a two-shell system, the larger (7.3’ in diameter) and more massive of which is located to the East and is centered on the Wd2 cluster, while the sec- ond shell (4.1’ in diameter), located to the West, surrounds WR20B. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | NOTE. — † These data were not used for the analysis. See the text for more details. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | photometric bands, from 3.6 µm to 24 µm (i.e. 3.6 µm, 4.5 µm, 5.6 µm, 8 µm and 24 µm), with a spatial resolution from 2” to 6”. We note that only partial data exist at 24µm, due to hard saturation in the bright core of the source. A comprehen- sive analysis of the GLIMPSE data for the region is provided in Whitney et al. (2004). This area of the sky was also covered by the Wide-field Infrared Survey Explorer (WISE, Wright et al. 2010) all-sky survey at 3.4, 4.6, 12 and 22 µm, with spatial resolution from 6” to 12”. As in the case of GLIMPSE, severe saturation of the core of RCW 49 occurs at 22 µm. Finally, RCW 49 was targeted by dedicated Herschel (Pilbratt et al. 2010) PACS (70 µm and 160 µm, at 6” and 12”, respectively) and SPIRE (250 µm, 350 µm, 500 µm, at 18”, 25” and 35”, respectively) Cycle 2 Open Time parallel mode observations (PI. R. Paladini). The analysis of these Herschel data will be the subject of a forthcoming publication (Paladini et al., in preparation). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | Noticeably, the brightest emission, at both radio and IR wavelengths, is found in correspondence of the ridge located between the two shells (see Figure 1). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | The paper is organized as follows. In Section 2 we describe the new ATCA continuum and radio recombination line ob- servations of the core of RCW 49. In Section 3 we provide details on the data reduction. In Section 4 we analyze this new data set. Finally, we discuss our results in Section 5, and present the conclusions in Section 6. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | At 5 GHz, in order to achieve the best uv-coverage, as well as the best S/N, we concatenated the data sets obtained in the 750 B and 1.5 C configurations. We also concatenated the 18 and 19 GHz data sets, as we do not expect any significant changes in the spectral index between these two frequencies. Hereafter we will only discuss results on the combined, final images. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | NOTE. — The parameters listed above for the 5 GHz map refer to the original data set, i.e. prior to filtering in the uv-plane (see Section 4.2). θmin and θmax indicate, respectively, the smallest and largest imaged angular scales. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 2.— ATCA 5 GHz map. The synthetic beam is shown in the bottom right corner. The map units are Jy/beam. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | all frequencies, with the exception of 5 GHz, we used a Briggs weighting, which allows a good match between noise and res- olution. The parameter robust was set to zero, to mini- mize sidelobe contributions from bright sources. At 5 GHz, we used a natural weighting scheme, in order to assign more weight to the shortest spatial frequencies, and achieve a better sensitivity at the largest angular scales. Deconvolution of the dirty images was performed by default with the Clark algo- rithm. Alternatively, the Högbom algorithm was used when the uv-coverage was too poor for imaging. In addition, a cor- rection for the primary beam was applied. Table 3 provides the synthetic beams, peak flux densities and noise of the final maps, shown in Figure 2, 3 and 4. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | same time, they could be ascribed, at least in part, to the sen- sitivity to different angular scales of the individual data sets. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | The expression above is an approximation of the Shaver et al. (1983) formula (see their Eq. (1)), with ν = 5 GHz and where we assume LTE conditions and that the fraction of ion- ized He is negligible (ne = nH+ ). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | Using the task viewer, we extracted a cumulative spec- trum within a box of the same size of the synthetic beam (∼ 7”) and centered at each position. In our system configuration we had a continuum bandwidth of 13 MHz, which allowed us to estimate the continuum level, TC, by fitting the data, minus the line, with a power-law model. A gaussian fit was applied to the continuum-subtracted data (see Figure 5). Both the con- tinuum and the line fits were carried out with the MATLAB fitting tool (cftool). The gaussian fit to the line gives the line peak temperature, TL, central velocity, V , and width to half-intensity, ∆V (see Table 4). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | We estimated the probability of a false detection. From Ta- ble 4, the statistical significance of the H109α line is ∼ 3σ in region A and B, and ∼ 2σ in region C, which corresponds to a probability of a spurious detection in a single velocity channel of, respectively, 1/370 and 1/50. Therefore, the prob- ability that at least one out of N velocity channels has a S/N of 3 (or 2) is given by 1 - (1 - 1/370)N (or 1 - (1 - 1/50)N). Since we apply the positivity condition10, the previous expres- sion becomes 1 - (1 - 1/740)N (or 1 - (1 - 1/100)N). We now consider the number of velocity channels that are effectively used for the detection of the line. First of all, the channels at the edge of bandwidth are discarded, as they are too noisy to provide any useful information. This operation brings the number N of channels from 512 to 400. In addition, we re- strict the search of the line to the velocity range +/- 30 km/s, which is the range of velocities compatible with the position of RCW 49 in the l-v diagram (see Figure 2 of Furukawa et al. 2009). This further decreases N from 400 to roughly 40 channels. Following these considerations, we finally estimate that the probability that the lines in region A and B are spuri- ous is 1 - (1 - 1/740)40 ∼ 0.05, while for region C is 1 - (1 - 1/100)40 ∼ 0.33. This calculation shows that our detections, especially in region C, are only marginal. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | with l the linear size of the emitting region. Table 4 pro- vides the derived values of Te, EM and ne by taking l = 1 pc. Uncertainties on these quantities were obtained by prop- agating the errors on TL, TC and ∆V . We notice a significant spread in Te for the three positions for which line extraction has been performed i.e, from ∼ 4000 K in region ’C’ to ∼ 16 000 K in region ’B’, indicating large variations in the gas physical conditions along the ridge. The average Te is 8468 ± 6326 K, which is consistent with the value reported by Caswell & Haynes (1987) of 7300 K. Our cumulative spec- tra are characterized by a mean velocity of 9.8 ± 10.7 km/s and by a mean velocity width of 8.2 ± 4.7 km/s, which is is in agreement with the H137β line observations by Benaglia et al. (2013). However, Caswell & Haynes (1987) and Church- well et al. (1974) detect line emission for the H109α and the He109α line, respectively, centered at 0 and -4 ± 1 km/s and with a line width of 46 and 50 km/s. We can ascribe this apparent discrepancy to the presence of several velocity com- ponents along the ridge, possibly reminiscent of independent structures which happened to collide at some point in time, as recently proposed by Furukawa et al. (2009) based on their NANTEN2 CO(J = 2-1) observations. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 3.— ATCA 19 GHz image of RCW49. The synthetic beam is shown in the bottom right corner. The beam PA is -12.3◦. The map units are Jy/beam. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 4.— ATCA 34 GHz image of RCW49. The synthetic beam is shown in the top left corner. The beam PA is 160.2◦. The map units are Jy/beam. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | We already mentioned that in the core of RCW 49 the Spitzer MIPS 24 µm data and the WISE 22 µm data are fully saturated. However, this is not an issue for our analysis. In fact, as discussed by various authors (e.g. Paladini et al. 2012; Everett & Churchwell 2010), in HII regions the emis- sion around 20 µm is likely associated to Big Grains rather than VSGs, contrary to what typically occurs in the Interstel- lar Medium (ISM). Therefore, to investigate the possible ex- istence of an IR-microwave correlation in RCW 49, we com- pared the ATCA continuum data at 5, 19 and 34 GHz with the IRAC 8 µm observations from the GLIMPSE survey. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | The comparison was performed both in the uv-plane and in real space. For this purpose, we used the CASA simulator tool which, at each observed ATCA frequency, allowed us to generate the corresponding visibilities for the IRAC 8 µm data set by taking into account the specific configura- tion of the interferometer. In particular, in order to obtain a perfect correspondence between observed and simulated uv- plane coverage, we used the combined sm.predict and sm.openfromms functions. As described in Section 2, the ATCA 5 GHz data set consists of a single pointing. In this case, to perform the simulation, a correction for the ATCA primary beam (∼ 570”) was applied by multiplying the IRAC 8 µm data with a Guassian profile, centered at the pointing phase center. On the contrary, the ATCA 19 and 34 GHz maps were obtained with a mosaic observing strategy. Therefore, a simulation was performed for each pointing and the correc- tion for the primary beam was applied using a Gaussian mask of 150” and 80”, respectively. Figure 6 shows the simulated vs observed visibilities at each frequency. In these plots, in order to minimize the correlations in the uv-plane, we binned the u, v cells in 10-min bins at 5 GHz and 1-min bins at 19 and 34 GHz, where the difference in bin width reflects the differ- ent integration time at the various frequencies. We computed the Pearson correlation coefficient11, ρ, and obtained: ρ8,5 = 0.936 ± 0.002, ρ8,19 = 0.848 ± 0.002, ρ8,34 = 0.783 ± 0.009, indicating that the correlation strength slightly weakens as we go from 5 to 19/34 GHz. Note that the correlation errors have been computed from the binned data, in linear scale, taking into account the number of data points (N−1/2) and their stan- dard deviation. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 5.— Cumulative continuum-subtracted spectra extracted from posi- tions ’A’ (top panel), ’B’ (middle panel) and ’C’ (bottom panel). The spectra have been smooothed using a 10-channel Hanning filter. The H109α line is centered on the LSR V ∼ 0 km/s. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 6.— The simulated IRAC 8 µm vs. the real ATCA visibilities. At 5 GHz (top panel), we notice a hint of a possibly residual correlation at high values of the visibilities. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 7.— Simulated IRAC 8 µm observations (white contours) overlaid on the observed ATCA data. From top to bottom: 5 GHz (750 B configuration), 19 GHz and 34 GHz (EW 352 configuration). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | Ideally, we should estimate the amount of microwave ex- cess in the ATCA data by extrapolating the low-frequency ob- servations (5 GHz) to the high-frequencies (19 and 34 GHz) and then by comparing the result with the CBI 31 GHz mea- surement. Since different interferometer configurations corre- spond to different amounts of flux loss, a correct approach to this problem would entail filtering the data in the uv-plane to assure that, at each frequency, we are sensitive to emission on the same angular scales. However, given the poor overlap in the visibility plane of the ATCA data set, if we followed this strategy we would remain with virtually no data to analyze. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | If we exclude the noisy pixels (i.e., those with an associated uncertainty greater than 0.1), the mean of the spectral index distribution is +0.20 ± +0.22. This is consistent (within 1.5σ) with optically thin free-free emission (α = -0.1), likely due to the ionizing flux from the OB stars in the Wd2 cluster. How- ever, the dispersion around the mean is quite significant. In particular, both the ridge and the West shell are characterized by positive spectral indices. Benaglia et al. (2013) have re- cently published two spectral index maps of RCW 49: the first one, derived from the combination of archival ATCA 1.4 and 2.4 GHz data12, covers both the ridge and the shells, while the second, generated from newly obtained 5.5 and 9.0 GHz data, covers only the ridge. The 1.4 – 2.4 GHz spectral index map is sensitive to angular scales up to several arcmins. The 5.5 – 9.0 GHz map is sensitive only up to scales of the order of 100”, that is a smaller scale than the maximum one which our 1.4 – 5.0 GHz map is sensitive to (∼ 3.4’). The spectral index map obtained from the lower frequency data (their Figure 5) appears to be fully compatible with optically thin free-free emission. This map also shows a possible contribution from synchrotron emission, especially along the Northern side of the ridge and in the West shell. Conversely, the spectral index map associated with the higher frequency data (their Figure 6) shows positive spectral indices. Indeed a large fraction of the ridge appears to be characterized by spectral indices greater than +0.3, in agreement with our analysis. Interestingly, val- ues of the order of +0.6 are typically associated with stellar winds (e.g., Panagia & Felli 1975) and, in general, a global distribution with rising spectral indices might corroborate the hypothesis, advanced by WU97, that the ridge is the product of the direct collision between the high-velocity winds pro- duced by WR20a and WR20b in the two shells or, alterna- tively, of a collision between the two shells before merging. Remarkably, this collision scenario could also explain the in- crease of dust temperature along the ridge noted by the same authors. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | FIG. 8.— Left: spectral index map obtained by combining the 5 GHz new observations of the core of RCW 49 with ATCA archival data at 1.4 GHz. The Sν ∼ να convention is used. Right: spectral index error map. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | From the analysis of the H109α data set described in Sec- tion 3.2, we estimated the emission measure and electron den- sity at three different positions along the ridge. We recall that no line emission was found, above the sensitivity level, at any other locations, although we should emphasize that the ther- mal continuum emission also encompasses only ∼ 1/20th of the 7.8(cid:48) × 5.6(cid:48) covered area (roughly 147 beams out of 2950). Table 4 shows that the average EM is ∼ 160 pc cm−6, with a corresponding ne of only ∼ 13 cm−3. These values are in- dicative of ionized gas denser than in typical ISM conditions, for which 0.03 < ne < 0.08 cm−3 (Haffner al. 2009), sup- porting the compressed-by shocks hypothesis formulated by WU97, yet this same gas appears to be much more diffuse than in HCHIIs. From the non-detection of H109α, we can set lower limits for the electron temperature and emission mea- sure through most of the 5 GHz map. For this purpose, we assume as representative of the non-detection region, Tc (cid:39) 60 mJy/beam and ∆V = 8.2 km/s. The latter is the average value for the H109α line in regions A, B and C. Setting the detection threshold at 3σ, where σ = 2.6 mJy/beam, and using Equation (1), (2), (3) and (4), we obtain Te > 9740 K, EM > 376 pc cm−6, and ne > 19.4 cm−3. Therefore our RRL mea- surements cannot completely rule out the presence of HCHIIs in the core of RCW 49. However, considering that the spectral index map derived in Section 4.2 shows no evidence of self- absorbed free-free emission (α ∼ 2), which would certainly characterize such dense condensations of ionized material, we confidently state that it is unlikely that the CBI detection can be attributed to a hidden population of HCIIs. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | and VSG spatial distributions, and that such a shift cannot be attributed to instrumental effects e.g., the CBI pointing error (up to 30”) or its low angular resolution. This finding alone suggests that either the detected AME is not due to spinning dust or, if the origin of the emission is indeed spinning dust, other factors (such as gas ions, etc.) might play a pivotal role in triggering this type of emission, as discussed in Tibbs et al. (2012). |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | with an average error of less than +0.1. These positive indeces are found mostly along the South side of the ridge and of the West shell. This rising of the spectral index distribution might suggest that other mechanisms, other than spinning dust, are responsible for the microwave excess observed by CBI, such as shocks due to a collision event either between the West and East shells or between the stellar winds generated by WR20b and WR20a, as speculated by WU97. Finally, the combined analysis of the spectral index map with the H109α recombi- nation line, marginally detected at three positions along the ridge, appears to suggest that the core of RCW 49 is likely not harboring very compact sources such as HCHII regions, thus not supporting an inverted free-free emission scenario as explanation for the CBI 31 GHz excess. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | We believe that this study has mostly shown the impor- tance of obtaining high angular resolution observations even for sources which have been studied for more than a century, like HII regions. These observations are in fact key in reveal- ing the complexity of physical mechanisms at work in these objects, in particular in very extended and evolved ones such as RCW 49. Without this detailed information it is difficult to interpret unambiguously the emission seen at relatively low angular resolution at microwave frequencies. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | Day, G. A., Caswell, J. L., Cooke, D. J., 1972, AuJPA, 25, 1 Deharveng, L., Zavagno, A. & Caplan, J., 2005, A&A, 433, 565 Dickinson, C., 2013, Advances in Astronomy, vol. 2013, 4 Dickinson, C., Davies, R. D., Allison, J. R., et al., 2009, ApJ, 690, 1585 Dickinson, C., Davies, R. D., Bronfman, L., et al., 2007, MNRAS, 379, 297 Dickinson, C., Casassus, S., Pineda, J. L., et al., 2006, ApJ, 643, 111 Draine, B. T., 2011, ApJ, 732, 100 Elmegreen, B. G., 1998, in ASP Conf. Ser. 148, ed. C. E. Woodward, J. M. |
Processed_Anomalous_Microwave_Emission_in_HII_regions:_is_it.txt | Tibbs, C.T., Paladini, R., Compiegne, M., et al., 2012, ApJ, 754, 17 Todorovic, M., Davies, R. D., Dickinson et al., 2010, MNRAS, 406, 1629 Vacca, W. D., Garmany, C. D., Shull, J. M., 1996, ApJ, 460, 914 van der Hucht, K. A., 2001, NewAR, 45, 135 Westerlund, B., 1960, Ark. Astron., 2, 419 Whiteoak, J. B. Z. & Uchida, K. I., 1997, A&A, 317, 563 Whiteoak, J. B. Z., Large, M. I., Cram, L. E., Piestryzynski, B., 1989, Proc. |
Processed_Controllability_and_data-driven_identification_of_.txt | Abstract— Nonlinear networked systems are of interest in several areas of research, such as multi-agent systems and social networks. In this paper, we examine the controllability of several classes of nonlinear networked dynamics on which the underlying graph admits negative weights. Such signed networks exhibit bipartite clustering when the underlying graph is structurally balanced. We show that structural balance is the key ingredient inducing uncontrollability when combined with a leader-node symmetry and a certain type of dynamical symmetry. We then examine the problem of extracting the bipartite structure of such graphs from data using Extended Dynamic Mode Decomposition to approximate the correspond- ing Koopman operator. |
Processed_Controllability_and_data-driven_identification_of_.txt | Consensus algorithms on networks admitting antagonistic interactions were considered by Altafini [11] and Pan et al. [24]. The graph-theoretic property of structural balance, used in the study of social networks, was identified as the property inducing bipartite consensus [25], [26], [27]. More- over, Clark et al. worked on the leader selection problem on signed networks [28]. Roy and Abaid discussed using antagonistic interactions to improve existing consensus-like algorithms [17]. |
Processed_Controllability_and_data-driven_identification_of_.txt | The generalization to nonlinear consensus algorithms has been studied in numerous settings. Behaviour of nonlin- ear consensus protocols was considered by Srivastava et al. [13]. The extension of these consensus protocols to signed networks was studied by Altafini [11]. Moreover, the generalization of symmetry arguments for controllability was examined by by Aguilar and Gharesifard [16]. |
Processed_Controllability_and_data-driven_identification_of_.txt | Networked dynamical systems are the cornerstone of many modern technologies, as well as the focus of scientific re- search in many disciplines. Some pertinent examples of net- worked systems are opinion dynamics [1], gene networks [2], flocking dynamics [3] and autonomous coordinated flight [4]. One of the most well-studied networked dynamical systems is the consensus protocol, which is used in many applications including Kalman filtering [5], [6], multi-agent systems [7], [8], the controls community has begun examining consensus on networks admitting both cooperative and antagonistic interactions [11], [12], [13], [14]. |
Processed_Controllability_and_data-driven_identification_of_.txt | Previous research in this direction has examined the con- trollability of consensus networks in both the linear case and non-linear generalizations of consensus [15], [16], [17]. The work by Rahmani et al. has shown that symmetries of networks characterized by graph automorphisms about leaders in the network cause uncontrollability [18]. This sym- metry technique was generalized by Chapman and Mesbahi, who showed that signed fractional automorphisms provide the necessary and sufficient conditions for characterizing controllability of consensus networks [19]. Further work examined methods of generating network topologies, for either performance improvements, as in [20] and [21], or those that are controllable for consensus, such as in [22] and [23]. |
Processed_Controllability_and_data-driven_identification_of_.txt | There has been a recent interest in applying data-driven methods to control of networks; one such approach has utilized the Koopman operator. The Koopman operator pro- vides a dynamical framework in which one considers the propagation of observables of the state, rather than the state itself. The Koopman operator is linear, even for a non- linear dynamical system, but the trade-off is that the vector space of observables is generally infinite-dimensional [29]. This formalism lends itself well to a data-driven approach, allowing one to approximate the Koopman operator by collecting data [30]. Research by Pan et al. has looked at identifying the bipartite structure of signed networks using data-driven methods [31], furthering work done by Facchetti et al. [32], and Harary and Kabell [33]. |
Processed_Controllability_and_data-driven_identification_of_.txt | The contributions of this paper are twofold. First, we show that the property of structural balance, when com- bined with symmetries in the underlying graph, as well as certain symmetries of the nonlinear dynamics, causes uncontrollability in the context of the accessibility problem. In particular, we consider the same network flows studied in [13], [16], [11]; however we extend the controllability analysis to signed dynamics. Secondly, we extend the bi- partite identification problem considered by Pan et al. in [31] to the case of signed nonlinear consensus networks. In particular, we use a Koopman operator-theoretic approach alongside Extended Dynamic Mode Decomposition (EDMD) to extract a ‘Koopman mode’ whose sign structure reveals the bipartite structure. |
Processed_Controllability_and_data-driven_identification_of_.txt | This research was supported by the U.S. Army Research Laboratory and the U.S. Army Research Office under contract number W911NF-13-1-0340, NSF grant SES-1541025 and AFOSR grant FA9550-16-1-0022. |
Processed_Controllability_and_data-driven_identification_of_.txt | II. MATHEMATICAL PRELIMINARIES A column vector with n elements is referred to as v ∈ Rn where vi or [v]i both represent the ith element in v. The square matrix N ∈ Rn×n is symmetric if N T = N . The identity matrix is denoted In. For w ∈ Rn the matrix diag(w) is an n × n matrix with w on its diagonal and zero elsewhere. We say that A is similar to B if there is an invertible matrix R such that R−1AR = B. The unit vector ei is the column vector with all zero entries except [ei]i = 1. The column vector of all ones is denoted as 1. The cardinality of a set S is denoted as |S|. The column space of a matrix M is denoted by R(M ). We denote the Moore-Penrose pseudoinverse of a matrix A as A†. The function h is even if h(−x) = h(x) and is odd if h(−x) = −h(x). The function f is of class Cr if the derivatives f, f ′, . . . , f (r) exist and are continuous. The function g ∈ C∞, otherwise called smooth, has derivatives of all orders. Let F : Rn → Rn be a vector field and let ϕ : Rn → Rn be a smooth mapping. Then, F is ϕ-invariant if (Dϕ(x))F (x) = F (ϕ(x)) for all x ∈ Rn with Dϕ(x) the Jacobian matrix of ϕ at x. Given a mapping γ : M → M, the fixed point set of γ is denoted by Fix(γ) = {x ∈ M|γ(x) = x}. |
Processed_Controllability_and_data-driven_identification_of_.txt | We follow the standard notation and conventions for graph theory applied to multi-agent systems and consensus, as in [15]. Below, we introduce some mathematics relating the ideas of symmetry and signed graphs that we will utilize in this paper. |
Processed_Controllability_and_data-driven_identification_of_.txt | An automorphism of the graph G is a permutation φ : V → V of its nodes such that φ(i)φ(j) ∈ E if and only if ij ∈ E. The permutation φ induces a mapping ϕ : Rn → Rn such that [ϕ(x)]i = xφ(i). Let the permutation matrix J be such that [J]ij = 1 if φ(i) = j and zero otherwise. Therefore, the permutation matrix J is simply the Jacobian matrix of ϕ, in that J = Dϕ. Thus, φ represents the automorphism of G if and only if JA(G) = A(G)J (see [15]). |
Processed_Controllability_and_data-driven_identification_of_.txt | In a similar way that we defined the functions φ and ϕ for the graph automorphism, consider the function g : V → V encoding the action of the gauge on the nodes.This induces a function g : Rn → Rn such that [g(x)]i = σixi. The gauge transformation is then the Jacobian of this function, in that Gt = Dg. |
Processed_Controllability_and_data-driven_identification_of_.txt | where F : Rn × Rm → Rn is a smooth mapping. The accessible set A(x0, T ) of the system (1) from x0 at time T is the set of all end-points θ(T ) where θ : [0, T ] → Rn is a trajectory of (1). The accessible set of (1) from x0 up to T is defined as A(x0, ≤ T ) := ∪0≤τ ≤T A(x0, τ ). Then, the nonlinear dynamical system (1) is said to be accessible from the initial point x0 if for every T > 0 the set A(x0, ≤ T ) contains a non-empty interior. |
Processed_Controllability_and_data-driven_identification_of_.txt | Consider the dynamical system in (1) without control. The Koopman operator K acts on functions of the state space (called observables) ψ by the action Kψ = ψ ◦ F . The function ϕ(x) is an eigenfunction of K corresponding to the eigenvalue µ if Kϕ(x) = eµtϕ(x). For an observable function g in the span of Koopman eigenfunctions, one ∞ can write g(x) = k=1 vkϕk(x), where the vk’s are the Koopman modes. For the case of full-state observable g(x) = x the states can be reconstructed as, x(t) = g(x(t)) = ∞ k=1 eµktϕk(x0)vk, where we refer to {µk, ϕk, vk} as the Koopman triple. A numerical approximation of Koopman P operator can be obtained by EDMD [30]. |
Processed_Controllability_and_data-driven_identification_of_.txt | In this paper, we tackle two problems regarding nonlinear signed consensus networks. In the first section, we extend the result in [16] to the signed networks. We examine the necessary conditions of uncontrollability in nonlinear signed network systems due to input and dynamics symmetry. In particular, we will show how the additional topological property of structural balance in signed networks plays a key role in driving uncontrollability. |
Processed_Controllability_and_data-driven_identification_of_.txt | In the second section of this paper, we show that a partic- ular Koopman mode from the EDMD approximation of the Koopman operator contains the sign structure corresponding to the bipartition in Lemma 1(5). This extends the work by Pan et al. who considered the equivalent problemn for linear signed consensus [31]. |
Processed_Controllability_and_data-driven_identification_of_.txt | To make the paper self-contained, we provide two main theorems from [16] which we use later to demonstrate uncontrollability. Theorem 1. Let G = (V, E) and F : Rn → Rn be a flow on G. Assume ϕ is a non-identity symmetry on F . Then, for any leader l, the leader-follower network flow on G induced by l is not accessible from the origin in Rn−1. Theorem 2. Let G = (V, E) and let F : Rn → Rn be the dynamics in any of (2)-(4). Also, assume ϕ be an automorphism of G. Then F is ϕ-invariant. |
Processed_Controllability_and_data-driven_identification_of_.txt | n x where x∗ f (t)dt → ∞ as |x| → ∞ and f (0) = 0; see [27] for properties of this class of functions. A subset R S0 ⊂ S of these functions that will be used later is the untranslated positive, infinite sector nonlinearities given by setting x∗ = 0 in the definition of S. The reason these classes of functions are interesting is that when combined with the dynamics introduced in (2)-(4), clustering occurs in a structurally balanced graph. This is summarized in the following theorem. Theorem 3. (Theorems 3 & 4 in [11]) Consider a graph G. Assume either the dynamics (2) with f ∈ S or the dynamics (3) with f ∈ S0 running on G. Then, limt→∞ x(t) = Gt1 if and only if G is structurally balanced 1 n (with gauge transformation Gt). |
Processed_Controllability_and_data-driven_identification_of_.txt | According to this theorem, for certain classes of functions, the dynamics will converge to two different clusters. These clusters are exactly those corresponding to the bipartite consensus condition in Lemma 1(5). |
Processed_Controllability_and_data-driven_identification_of_.txt | In the following subsections, we elaborate on the con- trollability of the dynamics (2)-(4) and show that a notion of symmetry about the input node, as well as structural balance, lead to uncontrollability. From [11] we know if the underlying signed graph G is structurally balanced, then there exists a gauge transformation Gt that acts as a similarity transformation on the adjacency matrix of G in that GtAs(G)Gt = A where A is the unsigned adjacency matrix. We will show that Gt defines a useful coordinate transformation that allows an immediate application of the uncontrollability test in [16]. |
Processed_Controllability_and_data-driven_identification_of_.txt | First, we discuss the notion of symmetry in signed graphs. Definition 1. Let ϕ be a non-identity automorphism on graph G. Suppose that this graph is structurally balanced, with gauge transformation Gt induced by the function g : Rn → Rn defined by [g(x)]i = σixi. Then, we define signed automorphism operator as ϕ′ = g ◦ ϕ ◦ g. Moreover, assume that J is the matrix representation of the permutation operator ϕ, in that J = Dϕ. Then, the analogous matrix J ′ = GtJGt is the matrix representation of the signed permutation operator ϕ′, in that J ′ = D(g ◦ ϕ ◦ g). By Definition 1, if ϕ(xi) = xr, then ϕ′(xi) = σiσrxr. For example, consider the graph in Figure 1. |
Processed_Controllability_and_data-driven_identification_of_.txt | Theorem 4. Consider a structurally balanced graph G with gauge transformation Gt and absolute nonlinear flow dy- namics (2). Further suppose G has a non-trivial signed automorphism ϕ′. Let f : Rn → R be a smooth odd function (for example, odd f ∈ S0 with f smooth). Then, for any vertex j ∈ Fix(ϕ′) chosen as the leader, the leader-follower network is not accessible from the origin in Rn−1. Moreover, the same results holds for smooth even functions f if ϕ′ preserves edge signs, in that sgn(aij ) = sgn(aφ(i)φ(j)). |
Processed_Controllability_and_data-driven_identification_of_.txt | which is an unsigned absolute nonlinear flow. Then, from Theorems 1 and 2 we conclude the system is inaccessible from the origin. |
Processed_Controllability_and_data-driven_identification_of_.txt | where r = φ(i) and l = φ(j) and the property f (σiσjx) = f (x) of even functions is used. |
Processed_Controllability_and_data-driven_identification_of_.txt | On the other hand, we know from the above identity that l∈Nr [f (xr) − sgn(arl)f (xl)] . |
Processed_Controllability_and_data-driven_identification_of_.txt | Fφ(i)(x) = Fr(x) = − Hence, F is ϕ′-invariant if sgn(aij ) = sgn(arl). The condition on even functions in Theorem 4 can be interpreted as an edge-sign symmetry of the graph, in that edge signs remain invariant under the signed automorphism. 2) Relative Nonlinear Flow: The main result of this section shows that structural balance and the existence of the non-trivial signed automorphism lead to the uncontrollability of the relative nonlinear flow. Theorem 5. Consider a structurally balanced graph G with gauge transformation Gt and relative nonlinear flow dynam- ics. Further suppose G has a non-trivial automorphism ϕ′. Let f : Rn → R be a smooth odd or even function (for example, odd f ∈ S0 with f smooth). Then, for any vertex j ∈ Fix(ϕ′) chosen as the leader, the leader-follower network is not accessible from the origin in Rn−1. |
Processed_Controllability_and_data-driven_identification_of_.txt | Proof: Let the same notations as in proof of Theorem 4 hold. We will show that both cases of odd and even functions f lead to ϕ′-invariance of the flow F and therefore the inaccessibility from the origin. |
Processed_Controllability_and_data-driven_identification_of_.txt | which is the unsigned relative nonlinear flow. Hence, F is ϕ′-invariant and inaccessibility from the origin follows from Theorems 1 and 2. |
Processed_Controllability_and_data-driven_identification_of_.txt | where the last display is equal to Fr(x), and we have used the fact that σiσj sgn(aij ) > 0 hence σiσj = sgn(aij) for all i and j ∈ Ni. |
Processed_Controllability_and_data-driven_identification_of_.txt | Theorem 6. Consider a structurally balanced graph G with gauge transformation Gt and disagreement nonlinear flow dynamics. Further suppose G has a non-trivial automorphism ϕ′. Let f : Rn → R be a smooth odd or even function (for example, odd f ∈ S0 with f smooth). Then, for any vertex j ∈ Fix(ϕ′) chosen as the leader, the leader-follower network is not accessible from the origin in Rn−1. |
Processed_Controllability_and_data-driven_identification_of_.txt | Proof: The proof is identical to that of Theorem 5. Remark 1. The analysis of this section demonstrates that for all of the three nonlinear flows (2)-(4) structural balance in addition to ϕ′-invariance leads to system uncontrollability for even and odd functions f . The only exception is when the absolute nonlinear flow f is even. In this case, an edge-sign symmetry condition is also required. |
Processed_Controllability_and_data-driven_identification_of_.txt | Since both α and ϕ1(x0) are constants, we can see that v1 ∝ Gt1, and the result follows. |
Processed_Controllability_and_data-driven_identification_of_.txt | In §VI-B, we show an example where we use EDMD to approximate the first mode of the Koopman operator to obtain the sign structure corresponding to the bipartite consensus. |
Processed_Controllability_and_data-driven_identification_of_.txt | In this section, we show some examples that highlight the necessity of combining ϕ-invariance, structural balance and leader-node symmetry for uncontrollability of signed networks. We refer the reader to [16] for similar examples in the unsigned case. We then show an example of using EDMD to obtain the bipartite consensus structure of a nonlinear flow on a structurally balanced graph. |
Processed_Controllability_and_data-driven_identification_of_.txt | Here we show an example of a network flow on a signed graph which has a leader-node symmetry. We show that in one case, the graph is structurally balanced, and the induced flow is hence uncontrollable. By altering the sign on a single edge, structural balance is lost and the resulting network flow is controllable. |
Processed_Controllability_and_data-driven_identification_of_.txt | Fig. 2: (a) Structurally balanced graph with a leader symme- try about node 4. (b) Structurally unbalanced graph with a leader symmetry about node 4. |
Processed_Controllability_and_data-driven_identification_of_.txt | In this subsection, we consider a numerical method to identify the bipartite consensus structure of nonlinear dy- namics on a structurally balanced graph, i.e. obtaining the bipartition of nodes in Lemma 1(5). We do this by exploiting Theorem 7, and numerically approximating the Koopman mode corresponding to the zero eigenvalue. |
Processed_Controllability_and_data-driven_identification_of_.txt | Consider the dynamics (3) with f (·) = sin(·), correspond- ing to the (signed) phase coupling of the Kuramoto dy- namics1, with an underlying structurally balanced oscillator network shown in Figure 3. |
Processed_Controllability_and_data-driven_identification_of_.txt | Fig. 3: Underlying structurally balanced signed graph for EDMD example. Dashed edges indicate negative edges. |
Processed_Controllability_and_data-driven_identification_of_.txt | We use EDMD to approximate the first Koopman mode corresponding to the zero eigenvalue. Due to space con- straints, we refer the reader to §2.2.3 of [30] for a detailed explanation of the EDMD algorithm. In particular, we use 6 a dictionary of functions of the form ψk = i=1 H(xi, ji) where H(xi, ji) is the jith order Hermite polynomial of xi. We use all such possible ψk for Hermite polynomials up to order 2. Simple combinatorics indicates that there are 729 such functions ψk for a 6-node graph; hence Nk = 729 in step (2) of §2.2.3 in [30]. |
Processed_Controllability_and_data-driven_identification_of_.txt | Fig. 4: Evolution of Dynamics 3 with f (·) = sin(·) on the graph in Figure 3. Left (red) shaded region corresponds to 1 and the right (blue) shaded region corresponds to ¯v3 ¯v2 1. |
Processed_Controllability_and_data-driven_identification_of_.txt | which all contain sign structure corresponding to the biparti- tion depicted in Figure 3. Despite not using all available data, the EDMD procedure was able to extract the bipartition well before the dynamics converged. This begs the question how early can we detect the bipartite structure of the underlying dynamics? We will address this question in future works. |
Processed_Controllability_and_data-driven_identification_of_.txt | In this paper, we examined the controllability of nonlinear flows on signed networks. In particular, we identified that the topological property of structural balance is the key ingredi- ent that when combined with a leader-node symmetry and ϕ- invariance of the flow dynamics, results in uncontrollability. We then looked at the task of identifying the bipartite consensus of certain classes of nonlinear network flows. We showed that the sign structure of the first Koopman mode corresponds to this bipartite consensus, and then used EDMD to numerically approximate this sign structure. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | This research's primary task is to inspect the affiliation between the implementation of technology and e- commerce’s success. It is imperative to study such an important relationship that directly impacts the rapid growth of Internet technology, new dimensions of e- services, and innovative measures that are necessary factors (e-commerce) operations. Despite most Saudi Arabia retailers being aware of technological advancements, existing research reveals several challenges that hinder the adoption of e-commerce strategies, including the cost of installation and training. The advantages of e- commerce are frequently shown in recent studies. Internet technologies development has narrowed the difference between trade and online traditional business grounds, with additional traditional markets moving to online platforms. The Saudi Arabia community has been recognized as a potential hub for advancing technology-based programs, particularly e- commerce, given a strong GDP per capita income and a 3.4 percent annual population growth. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | analyzing the current situation of e-commerce both in the country and in the city in particular. One of the outcomes would be the benefits of introducing e-commerce to modern SMEs, considering the characteristics of consumers in Riyadh and Saudi Arabia by utilizing a quantitative and qualitative mixed-method approach. Previous research in this area, such as "SMEs engagement with e-commerce, e- business, and e-marketing" [2] and "the impact of effective factors on the success of e-commerce in small-and-medium- sized companies" [3], has been conducted for promoting potential benefits of e-commerce technologies in today's these highly competitive markets. However, none of addressed the situation surrounding Saudi’s ecommerce development SMEs context. Due to different cultural and legal features, enterprises in Riyadh cannot simply use the experience of businesses from other countries because these approaches are not always suitable for the local environment. This study aims to fill the gaps in knowledge about the prospects and ways of developing e-commerce in Riyadh whole of Saudi Arabia. This study will contribute to bringing insightful knowledge for more effective development and strategic improvements of local SMEs, which are important to compete in both the domestic and foreign markets and support the economic development of the region. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Business technologies are the essential elements of everyday businesses and livings; computers, smartphones, and the Internet are many people's integral life parts. Effective electronic commerce (e-commerce) operation is crucial for small and medium enterprises (SMEs) in any market. The number of people who prefer to make purchases on the Internet is growing annually, where companies without e- commerce risk losing customers. Online purchases in Saudi Arabia grew from 18% to 40% between 2015 and 2017, significant [1] The growth of SMEs in developing countries is limited by poor e-commerce development compared to the developed countries using technology for a long. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | E-commerce is a significant technology in modern businesses, connection between business stakeholders electronically, allowing SMEs to achieve higher growth [4]. Businesses should follow e-commerce-trends, requiring Internet-enabled technologies to remain competitive. SMEs are rapidly adopting various e-business methods to improve performance [5]. E-commerce improves the value chain processes, enhancing distribution, and specifically streamlined transactions and communications. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | SMEs are undoubtedly one of the major economic growth forces in all countries [6]. Economic development in Saudi Arabia benefits from increased profitability and fair competition of SMEs across the region. SMEs should, therefore, consider adopting technologies that will result in favorable business outcomes. This research proposal will focus specifically on assessing the impact of technology and it’s usage ease and compatibility for e-commerce success in Saudi Arabia SMEs. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | This study aims at filling a gap caused by inadequate information in this area and enhance the study that links technology and e-commerce. The study conveys new prospects to Riyadh SMEs by determining the affiliation between technology embracing and effects on business’ results. Therefore, Riyadh SMEs managers can obtain a new understanding of how e-commerce technologies are important for their business practices. Moreover, new attributes will be studied and added to existing technology adoption theories, making recommendations to Riyadh SMEs. The e-commerce technologies' introduction into organizations, critical in today's highly competitive markets. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | within the limits of the adoption norm, there is always a moment when innovation reaches its critical mass, where further knowledge dissemination is impossible (Rogers, 1962). |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | the general strengthening of knowledge within the discipline of information systems, and in particular, an understanding of the process of introducing e-commerce and all the features of its implementation in the modern business and political environment of Saudi Arabia. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | More specifically, according to the Information systems success model (ISS) by [10], there are six most critical aspects of success for information systems, namely; information, system and services qualities, system usage intents, user gratification and overall system returns [10]. The aspects are interconnected with each other; hence, the flaws can affect the entire system. For the system to be successful, when introducing e-commerce, it is imperative to consider each factor. This study therefore examines these aspects of success in terms of e-commerce and complement existing knowledge in this area. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Besides, this study uses the technology acceptance model (TAM), which models the process as users accept and use it. According to this model, several factors directly or indirectly influence users' decisions about technology usage [11]. The model’s author identified two key factors, perceived utility and ease of use. Therefore, a particular system's users must improve their productivity and be easy to use [11]. The study will combine ISS and TAM to understand the main factors of e-commerce implementation, making it possible to give recommendations to enterprises. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | strategy. In terms of period, this study will be conducted within three months for collecting field data. The target population for this study will be managers of SMEs in Riyadh. The stratified sampling method will be used, and data will be collected through semi-structured questionnaires addressed to ICT experts in the enterprises. Such interviews the contain specific questions’ order wording is left to the interviewer’s direction [12]. This approach’s key feature is that the researcher can answer questions from participants, ask questions, and discuss issues of importance to them. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | to add more information regarding social media integration in Saudi Arabian businesses and how to widen their customer base. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Moreover, the study’s results should understand Saudi SMEs about what steps should be taken to gradually and efficiently in introduce e-commerce existing and new enterprises and avoid potential problems. It is vital to recognize that despite the willingness of Saudi Arabians to develop e-commerce in the country, SMEs must find a unique approach for their target audience so that potential customers have a positive online consumer experience and share it with others. Further investments in the region and economic stability are also critical. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | This study aims at finding out the benefits of adopting e-commerce technology in Saudi Arabian SMEs. The importance of technology in improving performance in modern businesses cannot be ignored. This study will fill a knowledge gap concerning e-commerce technologies in SMEs. While carrying out the study, technology will be used as the moderating variable, while its compatibility with the SMEs will be the mediating variable. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | will contribute technology. A. 2.2. E-commerce in the current business environment There have been numerous attentions on e-commerce in business, enabling the promotion of products hence better results. E-commerce has enhanced competition between online businesses. The difference between traditional and online retail has been reduced, as more traditional retails shops have switched to the use of online platforms [13]. Electronic devices use has also enhanced the use of e- commerce technologies. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | significant trends in e- commerce development insight into the state of it in the country and comparing it to other retailing technologies. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Sixty-five percent of Arabia's Saudi population has access to the Internet; online shoppers in the country have rapidly increased. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | It examines the features of advertising needed e- to promote after commerce implementing information technology. the It between relationship social networks and the development e- commerce as integral elements in modern markets. They offer a new solution for the cross-site cold-start recommendation. They examined how culture influences trust in e-commerce. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Advertising is undoubtedly one of the tools used to increase the any effectiveness process on the Internet. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | The authors determined that social networks are also factors the speed of affecting technology distribution and its relevance. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | Various statistical reports and studies show an increased integration of mobile and Internet technology in Riyadh, Saudi Arabia, over the last few years. This section illustrates studies that depict Saudi Arabia's demographics and resources and how they influence e-commerce adoption in the country. Riyadh's e-commerce market's growth is mainly attributed to the nation's GDP figures at one hundred and sixty-eight billion dollars making Saudi Arabia the largest economy throughout the Arab region [20]. Through its strong GDP per capita income and a 3.4 percent annual population growth (Al-Mosa, 2011), the Saudi Arabia community is a potential hub for advancing technology-based programs, particularly e-commerce. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | E-commerce in Saudi Arabia has gained significant attention leading to numerous studies and research on the topic. [21] surveyed online shoppers to understand the aspects of e-commerce acceptance in Riyadh. Furthermore, [16] reports that sixty-five percent of Saudi Arabia's population has access to the Internet. However, the report does not include details of the transactions made by related parties. According to a research report by [16], various factors such as user readiness and awareness, lack of trust, online modes of payment, internet privacy and safety, and credit card security are key inhibitors to the gradual adoption of online retailing. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | An efficient e-commerce infrastructure requires the combination of the necessary tools and technologies. The aspect ensures that the infrastructure have all the abilities to enhance the customer activities on a platform hence improving their satisfaction. One of the most implemented e- commerce technologies is mobile apps which are enormously suitable for consumers and retailers. The apps ensure that consumers do not have to visit stores physically which saves their time. Additionally, the applications bring more benefits to retailers and consumers. |
Processed_Impact_of_Business_technologies_on_the_success_of_.txt | There is a gap in this knowledge that will be covered in this research in the existing studies. There is no study that adequately the e-commerce developments (particularly for SMEs) in Saudi Arabia. There is a need to conduct more in-depth research to address the areas of e- commerce development in Riyadh and trends present among Saudi Arabia's SMEs. This will encourage more business to integrate modern technologies, that will aid them achieve greater business results by addressing dynamic competitive global-markets and contributing to the country's economic growth. [24] noted that government regulation and support were the major challenges in introducing e-commerce in Saudi Arabia. In turn, our study will focus solely on the technology adoption of SMEs to fill the knowledge gap. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.