The authors regret that the original manuscript reports incorrect values for some of the instances concerning the CP model in the case of UC-2 and UC-3. This affects only one subsection of the paper, that is Section 5.5.2. Despite this, the main conclusions of the analysis and of the paper do not change as the CP still delivers statistically inferior results w.r.t. the Simulated Annealing proposed in this work. The updated tables and figures are (Figures 1 and 2, Table 1) for completeness, we report also the case of UC-1, where the original manuscript was correct). The corrected gaps are as follows. For UC-2, the average gap is −6.23 % (±10.5); for UC-3, the average gap is −4.90 % (±5.76). With a total of 61 new upper bounds of UC-3 and 40 new upper bounds fo UC-2. The results reported in the additional material are correct, except for the following instances of the UC-2 CP method that should report –: 84, 88, 92, 94, 95, 98, 102, 103, 107, 110, 111, and 114. The authors would like to apologise for any inconvenience caused. [Figure presented] Figure 1. CD plots ranking SA w.r.t. state-of-the-art algorithms, namely ILP, CP, and LNS. The analysis in conducted by use case (UC) and by instance size (medium and large). [Figure presented] Figure 2. Comparison of SA w.r.t. the state-of-the-art results by use case (UC) and instance size. Table 1. Comparison of SA w.r.t. the state-of-the-art results by use case (UC) and instance size in terms of number of instances. [Figure presented]
Corrigendum to “Multi-neighborhood simulated annealing for the oven scheduling problem” [Comput. Oper. Res. 177 (2025) 106999] (Computers and Operations Research (2025) 177, (S0305054825000279), (10.1016/j.cor.2025.106999))
Da Ros F.
;Di Gaspero L.;
2025-01-01
Abstract
The authors regret that the original manuscript reports incorrect values for some of the instances concerning the CP model in the case of UC-2 and UC-3. This affects only one subsection of the paper, that is Section 5.5.2. Despite this, the main conclusions of the analysis and of the paper do not change as the CP still delivers statistically inferior results w.r.t. the Simulated Annealing proposed in this work. The updated tables and figures are (Figures 1 and 2, Table 1) for completeness, we report also the case of UC-1, where the original manuscript was correct). The corrected gaps are as follows. For UC-2, the average gap is −6.23 % (±10.5); for UC-3, the average gap is −4.90 % (±5.76). With a total of 61 new upper bounds of UC-3 and 40 new upper bounds fo UC-2. The results reported in the additional material are correct, except for the following instances of the UC-2 CP method that should report –: 84, 88, 92, 94, 95, 98, 102, 103, 107, 110, 111, and 114. The authors would like to apologise for any inconvenience caused. [Figure presented] Figure 1. CD plots ranking SA w.r.t. state-of-the-art algorithms, namely ILP, CP, and LNS. The analysis in conducted by use case (UC) and by instance size (medium and large). [Figure presented] Figure 2. Comparison of SA w.r.t. the state-of-the-art results by use case (UC) and instance size. Table 1. Comparison of SA w.r.t. the state-of-the-art results by use case (UC) and instance size in terms of number of instances. [Figure presented]I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
