Fig. 1 Illustration of cyber-physical interaction between digital twin and physical twin More about this image found in Illustration of cyber-physical interaction between digital twin and physica...

Fig. 2 Illustration of CDT and heterogeneous agents with embedded cognitive intelligence of perception, situational understanding, learning, and reasoning to optimize actions under disruptions More about this image found in Illustration of CDT and heterogeneous agents with embedded cognitive intell...

Fig. 3 Cognitive decision process for multi-objective production scheduling More about this image found in Cognitive decision process for multi-objective production scheduling

Fig. 4 Experimental configuration of cyber-physical twins in case studies including 20 machines (9 lathes, 4 milling machines, 3 drills, 3 welders, and 1 AGV charging station) and 6 AGVs for material handling More about this image found in Experimental configuration of cyber-physical twins in case studies includin...

Fig. 5 Experimental design for ( a ) evaluating role-specific impacts on KPIs (i.e., productivity, task transition, workload balance, and transport efficiency) and ( b ) benchmarking with conventional scheduling methods More about this image found in Experimental design for (a) evaluating role-specific impacts on KPIs (i.e.,...

Fig. 6 Impacts of role-specific objective weights α S ( Sender focus), α R ( Receiver focus), and α D ( Deliverer focus) on KPIs: ( a ) system productivity (measured as TIS) ( b ) task transition congestion (Queue-Out lengths) ( c ) workload balance (variance of Queue... More about this image found in Impacts of role-specific objective weights α S (Sender focus), ...

Fig. 7 Performance comparison of average TIS (time in system) across different scheduling approaches under varying job IATs and AGV counts, with ( a ) n A G V = 4 , ( b ) n A G V = 5 , and ( c ) n A G V = 6 , respectively More about this image found in Performance comparison of average TIS (time in system) across different sch...

Fig. 8 Performance comparison of KPIs—(1) task transition congestion ( Q ¯ O U T ), (2) workload balance ( W B ¯ ), and (3) transport efficiency ( T E I ¯ )—across different scheduling approaches under varying scenarios: ( a ) n A G V = 6 and... More about this image found in Performance comparison of KPIs—(1) task transition congestion ( Q ¯ ...

Fig. 9 Comparison of system throughput (number of finished jobs ( n J o b )) across different scheduling approaches under varying scenarios of job IATs and AGV counts, where ( a ) n A G V = 4 , ( b ) n A G V = 5 , and ( c ) n A G V = 6 More about this image found in Comparison of system throughput (number of finished jobs ( n J o b ...

Fig. 1 LPBF experimental arrangement with the pyrometry sensor data collection [ 18 ] More about this image found in LPBF experimental arrangement with the pyrometry sensor data collection [ 1...

Fig. 2 Methodology architecture for layer-wise emission prediction with explainable artificial intelligence where ( a ) is the non-interpretable flowchart and ( b ) is the explainable artificial intelligence framework More about this image found in Methodology architecture for layer-wise emission prediction with explainabl...

Fig. 3 Global mean absolute value plots for ( a ) SHAP, ( b ) IG, and ( c ) LIME More about this image found in Global mean absolute value plots for (a) SHAP, (b) IG, and (c) LIME

Fig. 4 Heat maps for boxes 6, 9, and 10 from ( a )–( c ) SHAP, ( d )–( f ) IG, and ( g )–( i ) LIME More about this image found in Heat maps for boxes 6, 9, and 10 from (a)–(c) SHAP, (d)–(f) IG, and (g)–(i)...