Addressing the challenges inherent in complex intelligent planning-specifically high-dimensional discrete decision spaces,strong constraint coupling,and the propensity of traditional single-layer Ising/Quadratic Uncon...Addressing the challenges inherent in complex intelligent planning-specifically high-dimensional discrete decision spaces,strong constraint coupling,and the propensity of traditional single-layer Ising/Quadratic Unconstrained Binary Optimization(QUBO)modeling to induce variable inflation and penalty imbalance-this paper proposes a hierarchical hybrid Ising simulation framework.This framework is investigated through its application to the problem of multi-agent task assignment and collaborative path planning.The proposed method models task assignment and visit sequencing as an outer-layer QUBO/Ising optimization problem,incorporating techniques such as continuous relaxation,momentum enhancement,noise perturbation,and feedback correction to bolster search capabilities.Concurrently,the inner layer handles path construction,conflict detection,and local repair,feeding the repair information back to the outer layer to establish a closed-loop optimization process.Experimental results demonstrate that this method outperforms comparative approaches across key metrics,including total cost,feasible solution rate,conflict control,and convergence stability.展开更多
Objective:To evaluate the dose uncertainty in stereotactic body radiation therapy induced by respiratory motion using a 4D dynamic dose(4DDD)reconstruction method.Methods:A retrospective analysis was conducted on five...Objective:To evaluate the dose uncertainty in stereotactic body radiation therapy induced by respiratory motion using a 4D dynamic dose(4DDD)reconstruction method.Methods:A retrospective analysis was conducted on five lung cancer patients who received static intensitymodulated radiation therapy.The 4DDD was constructed using beam delivery log files,four-dimensional computed tomography(4DCT)scans,and treatment plans.To evaluate the impact of respiratory motion,4DDD calculations were performed with 10 starting phases for each field.A total of 270 field doses were simulated and calculated.The differences between the cumulative volume histogram in whole-course treatment and the field doses'gamma passing rate(GPR)were compared.The correlations between plan complexity metrics and the dose deviation caused by respiratory motion were evaluated independently.The phase distributions of 398 subfields were calculated and evaluated for the influence of dose rate and breathing frequency.Results:The GPRs of all fields were different among various starting phases,with the highest range from 62.20% to 76.87% for 2 mm/3%GPR.The deviation of mean point dose was(5.42±5.21)%,and the deviation in the mean dose and D98% within the internal gross tumor volume were(0.97±0.71)% and(0.77±0.53)%,respectively.There was a significant correlation between the beam aperture-to-volume(BA2V)ratio and the average 2 mm/2%GPR(R?0.601,P<0.01).Lower dose rates led to a more homogeneous distribution of phases among subfields(t?44.100,P<0.001).Conclusions:Different beam starting phases had a limited impact on the overall treatment evaluation.However,the respiratory motion could be observed to induce dose deviations using the 4DDD reconstruction model,particularly for fields with small BA2V.展开更多
文摘Addressing the challenges inherent in complex intelligent planning-specifically high-dimensional discrete decision spaces,strong constraint coupling,and the propensity of traditional single-layer Ising/Quadratic Unconstrained Binary Optimization(QUBO)modeling to induce variable inflation and penalty imbalance-this paper proposes a hierarchical hybrid Ising simulation framework.This framework is investigated through its application to the problem of multi-agent task assignment and collaborative path planning.The proposed method models task assignment and visit sequencing as an outer-layer QUBO/Ising optimization problem,incorporating techniques such as continuous relaxation,momentum enhancement,noise perturbation,and feedback correction to bolster search capabilities.Concurrently,the inner layer handles path construction,conflict detection,and local repair,feeding the repair information back to the outer layer to establish a closed-loop optimization process.Experimental results demonstrate that this method outperforms comparative approaches across key metrics,including total cost,feasible solution rate,conflict control,and convergence stability.
基金supported by National Key R&D Program of China.(No.2016YFC0105311)National Natural Science Foundation of China(No.81803047)the Fundamental Research Funds for the Central Universities of China(No.2019kfyXKJC061).
文摘Objective:To evaluate the dose uncertainty in stereotactic body radiation therapy induced by respiratory motion using a 4D dynamic dose(4DDD)reconstruction method.Methods:A retrospective analysis was conducted on five lung cancer patients who received static intensitymodulated radiation therapy.The 4DDD was constructed using beam delivery log files,four-dimensional computed tomography(4DCT)scans,and treatment plans.To evaluate the impact of respiratory motion,4DDD calculations were performed with 10 starting phases for each field.A total of 270 field doses were simulated and calculated.The differences between the cumulative volume histogram in whole-course treatment and the field doses'gamma passing rate(GPR)were compared.The correlations between plan complexity metrics and the dose deviation caused by respiratory motion were evaluated independently.The phase distributions of 398 subfields were calculated and evaluated for the influence of dose rate and breathing frequency.Results:The GPRs of all fields were different among various starting phases,with the highest range from 62.20% to 76.87% for 2 mm/3%GPR.The deviation of mean point dose was(5.42±5.21)%,and the deviation in the mean dose and D98% within the internal gross tumor volume were(0.97±0.71)% and(0.77±0.53)%,respectively.There was a significant correlation between the beam aperture-to-volume(BA2V)ratio and the average 2 mm/2%GPR(R?0.601,P<0.01).Lower dose rates led to a more homogeneous distribution of phases among subfields(t?44.100,P<0.001).Conclusions:Different beam starting phases had a limited impact on the overall treatment evaluation.However,the respiratory motion could be observed to induce dose deviations using the 4DDD reconstruction model,particularly for fields with small BA2V.
