"... In PAGE 11: ... Identifying those mismatches does not give any feedback as to their origin. For instance, is the architecture or is the design wrong? Table3 shows a possible way of resolving all above mismatches without introducing new ones by raising the Inventory System to the same level as the Order Framework. We do not belief that the actual mismatch resolution should be done fully automatic.... ..."

"... In PAGE 20: ...2.1 Input Parameters Ten di erent numerical variants of the ve product problem, that are given in Table3 , were considered to study SMART performance. The variations among the parameter values are highlighted in the table containing the parameters.... In PAGE 24: ...Figure 4 shows the learning curve for the system #1 of Table3 , and also a plot showing the performance of the SMART agent (after learning) versus the AR and COR heuristics. Table 5 compares the average reward rate accrued by the policy learned by SMART versus that for the COR and AR heuristics for all 10 systems.... In PAGE 25: ... The two levels of the four factors that were considered are given in Table 6. Other parameters of the problem were kept the same as for system 1 of Table3 . At 90 per cent con dence level, only the xed cost of maintenance was found to be signi cant in a ecting the response (average reward).... In PAGE 26: ...5383 failures, and 3) total vacation time of the system. For system 2 of Table3 , maintenance cost was varied from a low value of $500 to a high value of $1200 (systems 2 through 10 of Table 3). Figure 5 shows the sensitivity plots.... ..."

"... In PAGE 9: ... If it is below the corresponding order-up-to level S, an order of size equal to shortfall of the corresponding inventory position from S is given to the supplier. Place Table3 here Note that one may come with more detailed inventory policies like considering both epochs as a decision epoch, and de ning distinct order-up-to levels based on di erent inventory positions. However, in this study we restrict ourselves to this simple policies.... ..."

"... In PAGE 16: ...investment to lagged fixed assets, It / Kt-1, where I denotes investment in fixed assets (the first and second columns of Table2 a) or in inventories (the third and fourth columns). We include several right-hand variables.... In PAGE 16: ...35 The right-hand variables are lagged because investment decisions typically take time to mature and are often implemented with delay. In the regressions displayed in Table2 a, we use the weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of monetary policy. We include the interest rate as a stand-alone regressor, and interacted with other variables.... In PAGE 17: ... The impact of the lagged interest rate on investment is substantial. Consider, for example, the regression displayed in the second column of Table2 a where Export share is not log-transformed. The coefficient on R(-1) indicates that for a firm that does not export ( Export share(-1) = 0), an increase of one percentage point in the interest rate on short-term credit (other things equal) reduces the ratio of investment to fixed assets by 1.... In PAGE 18: ...-1.4 plus 0.23*1.13), namely, the offsetting effect due to export intensity is about 20 percent. For investment in inventories (the fourth column of Table2 a), the order of magnitude of these coefficients is similar. A similar calculation can be performed for regressions where Export share is log-transformed yielding an offsetting effect which is on the same order of magnitude.... In PAGE 19: ...39 The magnitude of the coefficient on q is interpreted as follows. Consider the first column of Table2 a. If q increases by 1 (from the sample mean of 1.... In PAGE 19: ... As mentioned, these firms exhibit higher than average export intensity so this result is not surprising. Since dually listed firms are special in many respects, we removed them from the sample and repeated the regressions in Table2 a obtaining virtually identical results. We do not pursue this issue further.... In PAGE 20: ...Robustness Table2 b displays similar regressions with the dependent variable (investment) normalized by lagged sales income rather than by fixed assets. In columns numbered (1) we use a weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of the short-term interest rate.... In PAGE 20: ... We display specifications where the variables liquidity, leverage, and Govshare are included as stand-alone regressors and where they are interacted with R(-1). The results are similar to those in Table2 a. In particular, the coefficient on R(-1) is negative in all the columns, and significant in most, while the coefficient on R(-1) *Export share(-1) is positive and significant in all the columns.... In PAGE 20: ... It is evident from these additional regressions that the real investment of export-intensive firms is affected much less by domestic monetary policy, but pinning down the exact magnitude is difficult. In Table2 c we perform similar regressions using the nominal interest rate and inflation expectations as separate regressors. The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent.... In PAGE 20: ... The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent. In Table2 d we perform similar regressions using GLS where the data are weighted by log-assets. (We do not further correct the residuals for heteroskedasticity.... In PAGE 24: ...more strongly on the investment of firms that have less access to foreign currency denominated credit. As an additional check, we performed the regressions of Table2 a controlling for GDP growth (see Table 5b) obtaining virtually identical results and an insignificant coefficient on GDP growth. We also used aggregate consumption growth as an alternative control for demand obtaining the same results (not shown).... ..."

"... In PAGE 16: ...investment to lagged fixed assets, It / Kt-1, where I denotes investment in fixed assets (the first and second columns of Table2 a) or in inventories (the third and fourth columns). We include several right-hand variables.... In PAGE 16: ...35 The right-hand variables are lagged because investment decisions typically take time to mature and are often implemented with delay. In the regressions displayed in Table2 a, we use the weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of monetary policy. We include the interest rate as a stand-alone regressor, and interacted with other variables.... In PAGE 17: ... The impact of the lagged interest rate on investment is substantial. Consider, for example, the regression displayed in the second column of Table2 a where Export share is not log-transformed. The coefficient on R(-1) indicates that for a firm that does not export ( Export share(-1) = 0), an increase of one percentage point in the interest rate on short-term credit (other things equal) reduces the ratio of investment to fixed assets by 1.... In PAGE 18: ...-1.4 plus 0.23*1.13), namely, the offsetting effect due to export intensity is about 20 percent. For investment in inventories (the fourth column of Table2 a), the order of magnitude of these coefficients is similar. A similar calculation can be performed for regressions where Export share is log-transformed yielding an offsetting effect which is on the same order of magnitude.... In PAGE 19: ...39 The magnitude of the coefficient on q is interpreted as follows. Consider the first column of Table2 a. If q increases by 1 (from the sample mean of 1.... In PAGE 19: ... As mentioned, these firms exhibit higher than average export intensity so this result is not surprising. Since dually listed firms are special in many respects, we removed them from the sample and repeated the regressions in Table2 a obtaining virtually identical results. We do not pursue this issue further.... In PAGE 20: ...Robustness Table2 b displays similar regressions with the dependent variable (investment) normalized by lagged sales income rather than by fixed assets. In columns numbered (1) we use a weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of the short-term interest rate.... In PAGE 20: ... We display specifications where the variables liquidity, leverage, and Govshare are included as stand-alone regressors and where they are interacted with R(-1). The results are similar to those in Table2 a. In particular, the coefficient on R(-1) is negative in all the columns, and significant in most, while the coefficient on R(-1) *Export share(-1) is positive and significant in all the columns.... In PAGE 20: ... It is evident from these additional regressions that the real investment of export-intensive firms is affected much less by domestic monetary policy, but pinning down the exact magnitude is difficult. In Table2 c we perform similar regressions using the nominal interest rate and inflation expectations as separate regressors. The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent.... In PAGE 20: ... The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent. In Table2 d we perform similar regressions using GLS where the data are weighted by log-assets. (We do not further correct the residuals for heteroskedasticity.... In PAGE 24: ...more strongly on the investment of firms that have less access to foreign currency denominated credit. As an additional check, we performed the regressions of Table2 a controlling for GDP growth (see Table 5b) obtaining virtually identical results and an insignificant coefficient on GDP growth. We also used aggregate consumption growth as an alternative control for demand obtaining the same results (not shown).... ..."

"... In PAGE 12: ... In this case, it can be formulated as a single-period inventory stochastic problem with the mean of the demand depending on the shelf space. For non-perishable products, the problem can be addressed as a continuous review problem, which will constitute the second avenue for the proposed The optimal solution is to select items 1, 2, and 3 and to display them on all shelves with a total profit of 3381, see Table 7) It can be noticed from Table7 that, except for item 1 for which more units are displayed on the first shelf location, almost equal units (after rounding to the nearest integer values) are displayed on all shelves for the Table 7 Results of the optimization problem with 4-item and 4-location Item... ..."

"... In PAGE 16: ...investment to lagged fixed assets, It / Kt-1, where I denotes investment in fixed assets (the first and second columns of Table2 a) or in inventories (the third and fourth columns). We include several right-hand variables.... In PAGE 16: ...35 The right-hand variables are lagged because investment decisions typically take time to mature and are often implemented with delay. In the regressions displayed in Table2 a, we use the weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of monetary policy. We include the interest rate as a stand-alone regressor, and interacted with other variables.... In PAGE 17: ... The impact of the lagged interest rate on investment is substantial. Consider, for example, the regression displayed in the second column of Table2 a where Export share is not log-transformed. The coefficient on R(-1) indicates that for a firm that does not export ( Export share(-1) = 0), an increase of one percentage point in the interest rate on short-term credit (other things equal) reduces the ratio of investment to fixed assets by 1.... In PAGE 18: ...-1.4 plus 0.23*1.13), namely, the offsetting effect due to export intensity is about 20 percent. For investment in inventories (the fourth column of Table2 a), the order of magnitude of these coefficients is similar. A similar calculation can be performed for regressions where Export share is log-transformed yielding an offsetting effect which is on the same order of magnitude.... In PAGE 19: ...39 The magnitude of the coefficient on q is interpreted as follows. Consider the first column of Table2 a. If q increases by 1 (from the sample mean of 1.... In PAGE 19: ... As mentioned, these firms exhibit higher than average export intensity so this result is not surprising. Since dually listed firms are special in many respects, we removed them from the sample and repeated the regressions in Table2 a obtaining virtually identical results. We do not pursue this issue further.... In PAGE 20: ...Robustness Table2 b displays similar regressions with the dependent variable (investment) normalized by lagged sales income rather than by fixed assets. In columns numbered (1) we use a weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of the short-term interest rate.... In PAGE 20: ... We display specifications where the variables liquidity, leverage, and Govshare are included as stand-alone regressors and where they are interacted with R(-1). The results are similar to those in Table2 a. In particular, the coefficient on R(-1) is negative in all the columns, and significant in most, while the coefficient on R(-1) *Export share(-1) is positive and significant in all the columns.... In PAGE 20: ... It is evident from these additional regressions that the real investment of export-intensive firms is affected much less by domestic monetary policy, but pinning down the exact magnitude is difficult. In Table2 c we perform similar regressions using the nominal interest rate and inflation expectations as separate regressors. The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent.... In PAGE 20: ... The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent. In Table2 d we perform similar regressions using GLS where the data are weighted by log-assets. (We do not further correct the residuals for heteroskedasticity.... In PAGE 24: ...more strongly on the investment of firms that have less access to foreign currency denominated credit. As an additional check, we performed the regressions of Table2 a controlling for GDP growth (see Table 5b) obtaining virtually identical results and an insignificant coefficient on GDP growth. We also used aggregate consumption growth as an alternative control for demand obtaining the same results (not shown).... ..."

"... In PAGE 16: ...investment to lagged fixed assets, It / Kt-1, where I denotes investment in fixed assets (the first and second columns of Table2 a) or in inventories (the third and fourth columns). We include several right-hand variables.... In PAGE 16: ...35 The right-hand variables are lagged because investment decisions typically take time to mature and are often implemented with delay. In the regressions displayed in Table2 a, we use the weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of monetary policy. We include the interest rate as a stand-alone regressor, and interacted with other variables.... In PAGE 17: ... The impact of the lagged interest rate on investment is substantial. Consider, for example, the regression displayed in the second column of Table2 a where Export share is not log-transformed. The coefficient on R(-1) indicates that for a firm that does not export ( Export share(-1) = 0), an increase of one percentage point in the interest rate on short-term credit (other things equal) reduces the ratio of investment to fixed assets by 1.... In PAGE 18: ...-1.4 plus 0.23*1.13), namely, the offsetting effect due to export intensity is about 20 percent. For investment in inventories (the fourth column of Table2 a), the order of magnitude of these coefficients is similar. A similar calculation can be performed for regressions where Export share is log-transformed yielding an offsetting effect which is on the same order of magnitude.... In PAGE 19: ...39 The magnitude of the coefficient on q is interpreted as follows. Consider the first column of Table2 a. If q increases by 1 (from the sample mean of 1.... In PAGE 19: ... As mentioned, these firms exhibit higher than average export intensity so this result is not surprising. Since dually listed firms are special in many respects, we removed them from the sample and repeated the regressions in Table2 a obtaining virtually identical results. We do not pursue this issue further.... In PAGE 20: ...Robustness Table2 b displays similar regressions with the dependent variable (investment) normalized by lagged sales income rather than by fixed assets. In columns numbered (1) we use a weighted average (across banks) of the interest rate on non-indexed overdraft credit to businesses as a measure of the short-term interest rate.... In PAGE 20: ... We display specifications where the variables liquidity, leverage, and Govshare are included as stand-alone regressors and where they are interacted with R(-1). The results are similar to those in Table2 a. In particular, the coefficient on R(-1) is negative in all the columns, and significant in most, while the coefficient on R(-1) *Export share(-1) is positive and significant in all the columns.... In PAGE 20: ... It is evident from these additional regressions that the real investment of export-intensive firms is affected much less by domestic monetary policy, but pinning down the exact magnitude is difficult. In Table2 c we perform similar regressions using the nominal interest rate and inflation expectations as separate regressors. The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent.... In PAGE 20: ... The results are overall similar and in columns headed by (1), the coefficients on the nominal interest rate and on inflation expectations are virtually equal, suggesting that merging these variables into a single variable (the real interest rate) is roughly equivalent. In Table2 d we perform similar regressions using GLS where the data are weighted by log-assets. (We do not further correct the residuals for heteroskedasticity.... In PAGE 24: ...more strongly on the investment of firms that have less access to foreign currency denominated credit. As an additional check, we performed the regressions of Table2 a controlling for GDP growth (see Table 5b) obtaining virtually identical results and an insignificant coefficient on GDP growth. We also used aggregate consumption growth as an alternative control for demand obtaining the same results (not shown).... ..."

"... In PAGE 25: ...5 -2.9 Table5 : Comparison of the average reward rate incurred by SMART vs. AR and COR heuristics for the ve product production inventory problem.... ..."

"... In PAGE 3: ...Table 5: Recoverable automatic reductions Although the deletion of the proposed features results in a recoverable tagset, the resulting reduction in the size of the tagset is significant but not satisfactory. The first three items in Table5 do not involve any reduction in the tagset since they are the same in all tags of the given category (S, _ and _, respectively). Deleting the verbal root in adverbs brings about a most minimal decrease (1).... ..."