Experimental Results and Analysis
Experimental Results and Analysis
Embedded Files
Normalized Team Performance
Normalized Team Performance
The raw team mission scores (π) were normalized to create standards and consistent format (πππππ)
The raw team mission scores (π) were normalized to create standards and consistent format (πππππ)
πππππ = π / π
πππππ = π / π
A score table of the raw team performance (π)
A score table of the raw team performance (π)
A score table of the normalized team performance (πππππ)
A score table of the normalized team performance (πππππ)
Validation of the DRL-based affective workload allocation (Task A, D, F, and H wihtout AS)
Validation of the DRL-based affective workload allocation (Task A, D, F, and H wihtout AS)
<Details of rmANOVA results>
Comparison with Baseline (Task A) and our DRL models (Task D, F, H)
Comparison with Baseline (Task A) and our DRL models (Task D, F, H)
Performed rmANOVA with a post-hoc analysis (Bonferroni correction)
There is marginally significant at p < 0.05
Tasks with AWAA (Task D, F, and H) > Task AΒ
Average scores: Task A < Task D < Task HΒ < Task F
βΈ« Our model outperformed the fixed workload allocation method.
<Average of the obtained score in each set>
Analysis on obtained score in each set of baseline and our DRL models
Analysis on obtained score in each set of baseline and our DRL models
Performed rmANOVA with a post-hoc analysis (Bonferroni correction)
1st set: no significant different
2nd set:Β Team performance improved (marginally significant at p < .05)
3rd set: no significant different
βΈ« Our model can improve the team performance by changing the workload.
Comparison of Workload Allocation Methods (Task B, C, E, and G with AS)
Comparison of Workload Allocation Methods (Task B, C, E, and G with AS)
<Details of rmANOVA results>
Tasks with AS: Task B (Capability) vs. Task C (with IS) vs. Task E (with PS) vs. Task G (with Both IS and PS)
Performed rmANOVA with a post-hoc analysis (Bonferroni correction)
There are marginally significant differences (p < .05)Β
Average scores: Task C < Task E < Task BΒ < Task G
βΈ« Both PS and IS should be used for workload allocation method to achieve the highest team performance.
Interaction Effects Between Workload Allocation Methods (IS, PS, and AS)
Interaction Effects Between Workload Allocation Methods (IS, PS, and AS)
Performed two-way ANOVA to investigate the interaction effects
Β There is no significant difference at p = .05Β
But, there is a slight interaction effect
Β βΈ« Presence of AS is not affected on team performance & Presence PS or IS influence on team performance
Comparison of Equal vs. capability-based workload allocation (Task A vs. Task B)
Comparison of Equal vs. capability-based workload allocation (Task A vs. Task B)
Task A vs. Task B: Equal vs. capability-based workload allocation
There is no significant difference at p = .05
However, the mean of Task B is slightly higher than Task A;
βΈ« We observed that capability-based workload allocation better than fixed workload allocation methods.
Page updated
Google Sites
Report abuse