scholarly journals Evaluation of MRI sequences for quantitative T1 brain mapping

2017 ◽  
Author(s):  
P Tsialios ◽  
M Thrippleton ◽  
C Pernet
Keyword(s):  
T1 Mapping ◽  
Blood Perfusion ◽  
Quantitative Mri ◽  
T1 Relaxation Time ◽  
T1 Relaxation ◽  
Wide Range ◽  
Fitting Procedures ◽  
Speed Up ◽  
Mri Sequences ◽  
Lower Image Quality

AbstractT1 mapping constitutes a quantitative MRI technique finding significant application in brain imaging. It allows improved evaluation of contrast uptake, blood perfusion, volume, and provides more specific biomarkers of disease progression compared to conventional T1-weighted images. While there are many techniques for T1-mapping, there is also a wide range of reported T1-values in tissues, raising the issue of protocols’ reproducibility and standardization. The gold standard for obtaining T1-maps is based on acquiring IR-SE sequence. Widely used alternative sequences are IR-SE-EPI, VFA (DESPOT), DESPOT-HIFI and MP2RAGE that speed up scanning and fitting procedures. A custom MRI phantom was used to assess the reproducibility and accuracy of the different methods. All scans were performed using a 3T Siemens Prisma scanner. The acquired data were processed using two different codes. The main difference was observed for VFA (DESPOT) which grossly overestimated T1 relaxation time by 214 ms [CI: 126 270 ms] compared to the IR-SE sequence. MP2RAGE and DESPOT-HIFI sequences gave slightly shorter T1 than IR-SE (∼20 to 30ms) and can be considered as alternative and time-efficient methods for acquiring accurate T1 maps of the human brain, while IR-SE-EPI gave identical results, at a cost of a lower image quality.

Cardiac magnetic resonance using T1 mapping for assessment of late cancer therapeutics-related cardiotoxicity in childhood cancer survivors

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
R Panovsky ◽  
T Kepak ◽  
L Opatril ◽  
L Masarova ◽  
T Holecek ◽  
...  
Keyword(s):  
Heart Failure ◽  
Relaxation Time ◽  
T1 Mapping ◽  
Cardiac Involvement ◽  
Cancer Therapeutics ◽  
Childhood Cancer Survivors ◽  
Funding Source ◽  
T1 Relaxation Time ◽  
T1 Relaxation ◽  
Lv Ejection Fraction

Abstract Introduction Due to cardiotoxic treatment, childhood cancer survivors (CCS) are at 15-fold increased risk of developing chronic heart failure and are at 7-fold higher risk of premature death due to cardiac causes compared with the general population. Cardiac magnetic resonance (CMR) has the potential to detect early cardiac involvement with consecutive possibilities of preventive steps against the development of advanced stages of heart failure. So far, only a few studies using T1mapping in CCS monitoring have been published. Purpose This study aimed to assess early cardiac involvement in a population of CCS using T1 mapping. Methods One hundred five CCS of age 24,9±5,4 years were included, mean time since the end of cancer-therapeutics treatment was 12,2±5,8 years. One hundred of them underwent complete CMR examination at 1,5T scanner. Cine images for assessment of left ventricular (LV) volumetric and functional parameters, pre- and post-contrast Modified Look-Locker Inversion recovery (MOLLI) images were acquired for assessment of native T1 relaxation time and extracellular volume (ECV), and delayed postcontrast images for evaluation of late gadolinium enhancement (LGE). The measured parameters were compared between CCS patients and 50 healthy controls. Results CCS patients had not enlarged LV (end-diastolic volume 128±30ml vs 124±30ml, p=NS) and normal, although lower systolic LV function than the controls – LV ejection fraction 59±6% vs 67±5% (p<0,05). In CCS group, only 3 (3%) patients had LV ejection fraction <50%, four (4%) patients had regional LV hypokinesia, and small non-ischemic LGE was found in 4 (4%) patients. Mean native T1 relaxation time was not prolonged - 987±31 msec vs 986±24msec (p=NS), and mean ECV value was not increased – 24,6±5,3% vs 23,8±2,3% (p=NS). Conclusion In this study, the late cancer therapeutics-related cardiotoxicity was quite low. Native and postcontrast T1mapping did not show any significant subclinical myocardial involvement. Funding Acknowledgement Type of funding source: Public grant(s) – EU funding. Main funding source(s): Supported by the European Regional Development Fund - Project ENOCH (no. CZ.02.1.01/0.0/0.0/16_019/0000868)

scholarly journals Value of T1 Mapping on Gadoxetic acid-enhanced MRI for Microvascular Invasion of Hepatocellular Carcinoma: A Retrospective Study

2020 ◽  
Author(s):  
chenyi Rao ◽  
Xinquan Wang ◽  
Minda Li ◽  
Guofeng Zhou ◽  
Hongmei Gu
Keyword(s):  
Hepatocellular Carcinoma ◽  
Relaxation Time ◽  
T1 Mapping ◽  
Operating Characteristic ◽  
Reduction Rate ◽  
Gadoxetic Acid ◽  
T1 Relaxation Time ◽  
T1 Relaxation ◽  
Enhanced Mri ◽  
Adc Values

Abstract Background: To evaluate the utility of non-invasive parameters driving from T1 mapping on gadoxetic acid-enhanced MRI for predicting microvascular invasion (MVI) of hepatocellular carcinoma (HCC) compared with diffusion-weighted imaging (DWI).Methods: A total of 94 patients with single HCC undergoing partial hepatectomy was included in the retrospective study, who underwent preoperative gadoxetic acid-enhanced MRI combined with DWI and T1 mapping. Parameters including precontrast, postcontrast and reduction rate of T1 relaxation time and ADC values were measured for differentiating MVI-positive HCCs (n=38) from MVI-negative HCCs (n=56). The receiver operating characteristic curve (ROC) was analyzed to compare the diagnostic performance of the calculated parameters.Results: The mean value of postcontrast T1 relaxation time were significantly higher in MVI-positive HCCs that MVI-negative HCCs (621.0 vs. 536.5, P <0.001). MVI-positive HCCs demonstrated significantly lower reduction rates of T1 relaxation time and lower ADC values than MVI-negative HCCs (39.4% vs 49.9, P<0.001; 1.495×10-3mm2/s vs 1.620×10-3mm2/s, P=0.003, respectively). The area under receiver operating characteristic curves were 0.587, 0.728, 0.824 and 0.690 for precontrast, postcontrast, reduction rate of T1 relaxation time and ADC, respectively. The reduction rate of T1 relaxation time was the most reliable feature with sensitivity, specificity and accuracy of the cut-off value (44.9%) of 79.0%, 73.2%, 75.5%, respectively.Conclusions: Reduction rate of T1 relaxation time on gadoxetic acid-enhanced MRI holds promise for evaluating MVI status of HCC.

scholarly journals P1604 T1 mapping in asymptomatic patients with extracardiac sarcoidosis

2020 ◽  
Vol 21 (Supplement_1) ◽  
Author(s):  
R Panovsky ◽  
M Doubkova ◽  
T Holecek ◽  
J Machal ◽  
V Feitova ◽  
...  
Keyword(s):  
Relaxation Time ◽  
Respiratory Tract ◽  
T1 Mapping ◽  
Left Ventricular ◽  
T1 Relaxation Time ◽  
Post Contrast ◽  
Native T1 ◽  
T1 Relaxation ◽  
Asymptomatic Patients ◽  
Molli Sequence

Abstract Funding Acknowledgements LQ1605 from the National Program of Sustainability II (MEYS CR) Introduction Sarcoidosis is a systemic granulomatous disease affecting in particular the respiratory tract. Estimated 5% of these patients have clinical symptoms of heart involvement. Real number of patients with cardiac sarcoidosis is thought to be higher (around 20-30%), because the granulomatous process is sometimes asymptomatic. Cardiac magnetic resonance (CMR) including T1 relaxation time measurement could potentially detect early asymptomatic stadia of sarcoidosis of the heart. Purpose The aim of this study was to assess T1 mapping in detection of early cardiac involvement in asymptomatic patients with sarcoidosis. Methods One hundred twenty patients with sarcoidosis of the respiratory tract and/or extrapulmonary sarcoidosis and without any heart disease history were included. One hundred thirteen of them underwent CMR examination at 3,0 MR scanner. Cine images for assessment of left ventricular (LV) volumetric and functional parameters, and pre- and post-contrast Saturation method using adaptive recovery times for cardiac T1 mapping (SMART1map) and Modified Look-Locker Inversion recovery (MOLLI) images were acquired for assessment of native T1 relaxation time and extracellular volume (ECV). The measured parameters were compared between sarcoidosis patients and 22 healthy controls. Results Sarcoidosis patients had not enlarged LV (end-diastolic volume 119 ± 24ml vs 97 ± 20ml, p = NS) and normal global and regional systolic LV function – LV ejection fraction (EF) 65 ± 5% vs 66 ± 7% (p = NS). Mean native T1 relaxation time was not prolonged - 1464 ± 93 msec vs 1482 ± 88msec (p = NS) measured by SMART1map and 1317 ± 60 msec vs 1313 ± 83msec (p = NS) measured using MOLLI sequence. Similarly, mean ECV value was not increased - 16,3 ± 3,4% vs 17,9 ± 3,7% (p = NS) measured by SMART1map and 30,9 ± 2,9msec vs 30,7 ± 5,0% (p = NS) measured using MOLLI sequence. Conclusion Myocardial native T1 relaxation time was not prolonged and ECV was not increased in asymptomatic patients with extracardiac sarcoidosis.

scholarly journals Longer sustained Fontan circulation is associated with prolonged hepatic T1 relaxation time on T1 mapping

2021 ◽  
Vol 22 (Supplement_2) ◽  
Author(s):  
A Yanovskiy ◽  
T Ojala ◽  
R Kivisaari ◽  
L Martelius
Keyword(s):  
Relaxation Time ◽  
Liver Disease ◽  
T1 Mapping ◽  
Single Ventricle ◽  
Venous Pressure ◽  
Relaxation Times ◽  
Hemodynamic Parameters ◽  
T1 Relaxation Time ◽  
T1 Relaxation ◽  
Systemic Ventricle

Abstract Funding Acknowledgements Type of funding sources: None. Background Patients with single ventricle defects may develop Fontan-associated liver disease. T1 mapping has been successfully used for evaluating chronic liver disease in adults. Liver T1 mapping has been also studied in the pediatric patients with single ventricles, and these patients show higher T1 relaxation times compared to the healthy controls. Purpose Our objective was to study the relationship between the cardiac MRI (CMR) T1 mapping relaxation time of the liver and 1) CMR derived hemodynamic parameters, 2) peripheral venous pressure (PVP) measured from a cubital cannula 3) systemic ventricle morphology [LV vs. RV], 4) the age of patient, and 5) alanine transaminase (P-ALAT) levels. Methods This retrospective study included 46 patients with functional single ventricle, which underwent routine CMR at our hospital. Table 1 shows demographic and clinical data of the study population. Statistical analysis were performed with IBM SPSS Statistics v.25 software using independent samples t test, Mann-Whitney U-test or Pearson correlation as appropriate. A p-value less than 0.05 was considered significant. Results The average T1 relaxation time of the liver was longer in patients with RV morphology (p = 0.004). There was a significant moderate positive correlation between the age of the patients and hepatic T1 relaxation time (r = 0.45, p = 0.002), and between hepatic T1 relaxation time and P-ALAT levels (r = 0.5, p = 0.016) (Fig.1).  No significant correlations were detected between the T1 times of the liver and hemodynamic parameters of the heart (all tested parameters are listed in the Table1). Ejection fraction and PVP showed a non-significant weak correlation with a hepatic T1 relaxation times (r=-0.3, p = 0,056 and r = 0.3, p = 0,070, respectively). Conclusions  T1 mapping times of the liver may reflect Fontan-associated liver disease. We observed connections between the hepatic T1 relaxation times and 1) patients age, 2) systemic ventricle morphology and 3) P-ALAT levels.

scholarly journals Value of T1 Mapping on Gadoxetic acid-enhanced MRI for Microvascular Invasion of Hepatocellular Carcinoma: A Retrospective Study

2020 ◽  
Author(s):  
chenyi Rao ◽  
Xinquan Wang ◽  
Minda Li ◽  
Guofeng Zhou ◽  
Hongmei Gu
Keyword(s):  
Relaxation Time ◽  
T1 Mapping ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Reduction Rate ◽  
Gadoxetic Acid ◽  
T1 Relaxation Time ◽  
Adc Value ◽  
T1 Relaxation ◽  
Enhanced Mri

Abstract Background: To evaluate the utility of non-invasive parameters derived from T1 mapping and diffusion-weighted imaging (DWI) on gadoxetic acid-enhanced MRI for predicting microvascular invasion (MVI) of hepatocellular carcinoma (HCC).Methods: A total of 94 patients with single HCC undergoing partial hepatectomy was analyzed in this retrospective study. Preoperative T1 mapping and DWI on gadoxetic acid-enhanced MRI was performed. The parameters including precontrast, postcontrast and reduction rate of T1 relaxation time and apparent diffusion coefficient (ADC) values were measured for differentiating MVI-positive HCCs (n=38) from MVI-negative HCCs (n=56). The receiver operating characteristic curve (ROC) was analyzed to compare the diagnostic performance of the calculated parameters.Results: MVI-positive HCCs demonstrated a significantly lower reduction rate of T1 relaxation time than that of MVI-negative HCCs (39.4% vs 49.9, P<0.001). The areas under receiver operating characteristic curve (AUC) were 0.587, 0.728, 0.824, 0,690 and 0.862 for the precontrast, postcontrast, reduction rate of T1 relaxation time, ADC and the combination of reduction rate and ADC, respectively. The cut-off value of the reduction rate and ADC calculated through maximal Youden index in ROC analyses was 44.9% and 1553.5 s/mm2. To achieve a better diagnostic performance, the criteria of combining the reduction rate lower than 44.9% and the ADC value lower than 1553.5 s/mm2 was proposed with a high specificity of 91.8% and accuracy of 80.9%.Conclusions: The proposed criteria of combining the reduction rate of T1 relaxation time lower than 44.9% and the ADC value lower than 1553.5 s/mm2 on gadoxetic acid-enhanced MRI holds promise for evaluating MVI status of HCC.

scholarly journals Value of T1 Mapping on Gadoxetic acid-enhanced MRI for Microvascular Invasion of Hepatocellular Carcinoma: A Retrospective Study

2020 ◽  
Author(s):  
Chenyi Rao ◽  
Xinquan Wang ◽  
Minda Li ◽  
Guofeng Zhou ◽  
Hongmei Gu
Keyword(s):  
Relaxation Time ◽  
T1 Mapping ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Reduction Rate ◽  
Gadoxetic Acid ◽  
T1 Relaxation Time ◽  
Adc Value ◽  
T1 Relaxation ◽  
Enhanced Mri

Abstract Background: To evaluate the utility of non-invasive parameters driving from T1 mapping and diffusion-weighted imaging (DWI) on gadoxetic acid-enhanced MRI for predicting microvascular invasion (MVI) of hepatocellular carcinoma (HCC). Methods: A total of 94 patients with single HCC undergoing partial hepatectomy was analyzed in the retrospective study. Preoperative T1 mapping and DWI on gadoxetic acid-enhanced MRI was performed. The parameters including precontrast, postcontrast and reduction rate of T1 relaxation time and apparent diffusion coefficient (ADC) values were measured for differentiating MVI-positive HCCs (n=38) from MVI-negative HCCs (n=56). The receiver operating characteristic curve (ROC) was analyzed to compare the diagnostic performance of the calculated parameters. Results: MVI-positive HCCs demonstrated a significantly lower reduction rate of T1 relaxation time than that of MVI-negative HCCs (39.4% vs 49.9, P<0.001). The areas under receiver operating characteristic curve (AUC) were 0.587, 0.728, 0.824 and 0,690 for precontrast, postcontrast, reduction rate of T1 relaxation time and ADC, respectively. The cut-off value of the reduction rate and ADC calculated though maximal Youden index in ROC analyses was 44.9% and 1553.5 s/mm2. To achieve a better diagnostic performance, the criteria that combining the reduction rate lower than 44.9% and the ADC value lower than 1553.5 s/mm2 was proposed with a high specificity of 91.8% and accuracy of 80.9%. Conclusions: The criteria that combining the reduction rate of T1 relaxation time lower than 44.9% and the ADC value lower than 1553.5 s/mm2 on gadoxetic acid-enhanced MRI holds promise for evaluating MVI status of HCC.

scholarly journals Value of T1 Mapping on Gadoxetic acid-enhanced MRI for Microvascular Invasion of Hepatocellular Carcinoma: A Retrospective Study

2020 ◽  
Author(s):  
chenyi Rao ◽  
Xinquan Wang ◽  
Minda Li ◽  
Guofeng Zhou ◽  
Hongmei Gu
Keyword(s):  
Relaxation Time ◽  
T1 Mapping ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Reduction Rate ◽  
Gadoxetic Acid ◽  
T1 Relaxation Time ◽  
Adc Value ◽  
T1 Relaxation ◽  
Enhanced Mri

Abstract Background: To evaluate the utility of non-invasive parameters derived from T1 mapping and diffusion-weighted imaging (DWI) on gadoxetic acid-enhanced MRI for predicting microvascular invasion (MVI) of hepatocellular carcinoma (HCC). Methods: A total of 94 patients with single HCC undergoing partial hepatectomy was analyzed in this retrospective study. Preoperative T1 mapping and DWI on gadoxetic acid-enhanced MRI was performed. The parameters including precontrast, postcontrast and reduction rate of T1 relaxation time and apparent diffusion coefficient (ADC) values were measured for differentiating MVI-positive HCCs (n=38) from MVI-negative HCCs (n=56). The receiver operating characteristic curve (ROC) was analyzed to compare the diagnostic performance of the calculated parameters. Results: MVI-positive HCCs demonstrated a significantly lower reduction rate of T1 relaxation time than that of MVI-negative HCCs (39.4% vs 49.9, P <0.001). The areas under receiver operating characteristic curve (AUC) were 0.587, 0.728, 0.824, 0,690 and 0.862 for the precontrast, postcontrast, reduction rate of T1 relaxation time, ADC and the combination of reduction rate and ADC, respectively. The cut-off value of the reduction rate and ADC calculated through maximal Youden index in ROC analyses was 44.9% and 1553.5 s/mm 2 . To achieve a better diagnostic performance, the criteria of combining the reduction rate lower than 44.9% and the ADC value lower than 1553.5 s/mm 2 was proposed with a high specificity of 91.8% and accuracy of 80.9%. Conclusions: The proposed criteria of combining the reduction rate of T1 relaxation time lower than 44.9% and the ADC value lower than 1553.5 s/mm 2 on gadoxetic acid-enhanced MRI holds promise for evaluating MVI status of HCC.

Abstract 13596: Magnetic Resonance - Derived Pre-contrast T1 Relaxation Time is the Accurate Marker of Diffuse Myocardial Fibrosis in Severe Aortic Valve Disease: A Comparison With Left Ventricular Myocardial Biopsy

Circulation ◽  
2015 ◽  
Vol 132 (suppl_3) ◽  
Author(s):  
Radka Kockova ◽  
Petr Kacer ◽  
Jan Pirk ◽  
Jiri Maly ◽  
Martina Vsianska ◽  
...  
Keyword(s):  
Relaxation Time ◽  
Myocardial Fibrosis ◽  
T1 Mapping ◽  
Interventricular Septum ◽  
Left Ventricular ◽  
Diffuse Myocardial Fibrosis ◽  
T1 Relaxation Time ◽  
Native T1 ◽  
T1 Relaxation ◽  
Myocardial Collagen

Introduction: Diffuse myocardial fibrosis (DFM) is the major mechanism in the pathophysiology of the aortic stenosis and its complications. DMF is detectable by magnetic resonance imaging (MRI) using the T1 mapping technique. Hypothesis: The MRI derived native T1 relaxation time and myocardial extracellular volume fraction (ECV) will be significantly related to the extent of DMF et targeted myocardial left ventricular (LV) biopsy. Methods: The study population consisted of 40 consecutive patients (age 63±8y, 65% males) undergoing surgery for severe aortic stenosis (77.5%), aortic root dilatation (7.5%) or valve regurgitation (15%). All patients underwent MRI-derived T1 mapping and 2D-, 3D speckle tracking-derived strain analysis prior to surgery. The T1 relaxation time was assessed in basal interventricular septum pre and 10 min post contrast administration using the modified Look-Locker Inversion recovery sequence. A LV myocardial biopsy specimen was obtained during surgery from basal interventricular septum under the guidance of the MRI operator to assure spatial concordance with the MRI assessment. The percentage of myocardial collagen was quantified as a ratio of Picrosirius Red-positive area over total sample area using the Image J. Results: The average percentage of myocardial collagen was 22 ± 14.8 %. The average native T1 relaxation time and ECV was 1010 ± 48 ms and 0.288 ± 0.055, respectively. Both native T1 relaxation time with cutoff value of ≥ 1010 ms (Ss=90%, Sp=73%, AUC =0.82) and ECV with cutoff value of ≥ 0.315 (Ss=80%, Sp=90%, AUC =0.85) showed high accuracy to identify extensive (> 30%) myocardial collagen content (Figure 1A, 1B). The native T1 mapping showed significant correlation with LV mass, 2D and 3D global longitudinal strain (all p<0.05) while the ECV did not (p=NS). Conclusions: Native T1 relaxation time is the accurate marker of diffuse myocardial fibrosis with the significant relationship with LV morphology and myocardial function.

T1 Relaxation Time of Achilles Tendon at 3 Tesla with Special Reference to Relevant Clinical Score: A Preliminary Study

2020 ◽  
Vol 16 (2) ◽  
pp. 164-173
Author(s):  
Zeineb Tbini ◽  
Mokhtar Mars ◽  
Mouna Bouaziz
Keyword(s):  
Relaxation Time ◽  
Achilles Tendon ◽  
Clinical Score ◽  
3 Tesla ◽  
Acquisition Time ◽  
Short Term ◽  
T1 Relaxation Time ◽  
T1 Relaxation ◽  
Spin Echo Sequence ◽  
Significant Difference

Purpose: The purpose of this study was to investigate T1 relaxation time of the human Achilles tendon, to test its short-term repeatability as well as the minimal detectable change, and to assess the extent that correlate with clinical symptoms. Methods: Twenty asymptomatic volunteers and eighteen patients with clinically and sonographically confirmed tendinopathy were scanned for ankle using a 3 Tesla (T) MR scanner. T1 maps were calculated from a variable flip angle gradient echo Ultra-short echo time sequence (VFA-GE UTE) and inversion recovery spin echo sequence (IR-SE) using a self-developed matlab algorithm in three regions of interest of Achilles Tendon (AT). Signal to Noise Ratio (SNR) between the two sequences was evaluated. INTRA-class Correlation Coefficient (ICC), Coefficient of Variation (CV) and the Least Significant Change (LSC) were calculated, to test short-term repeatability of T1. Subjects were assessed by the VISA-A clinical score. P values less than 0.005 were considered statistically significant. Results: Mean T1 values were 427.09 ± 53.37 ms and 528.70 ± 103.50 ms using IR-SE sequence and 575.43 ± 110.60 ms and 875.81 ± 425.77 ms with VFA-GE UTE sequence in the whole AT for volunteers and patients, respectively. : T1 values showed a significant difference between volunteers and patients (P=0.001). Regional variation of T1 in healthy and tendinopathic AT were greater for VFA-GE UTE sequence than for IR-SE sequence. VFA-GE UTE sequence showed clearly higher SNR compared to IR-SE sequence. Short-term repeatability of T1 values for volunteers showed an LSC of 22% and 14% for IR-SE sequence and VFA-GE UTE sequence, respectively. For patients, LSC was 14% and 5% for IR-SE sequence and VFA-GE UTE sequence, respectively. There was no correlation between T1 and VISA-A clinical score (p>0.005). Conclusion: VFA-GE UTE sequence used for T1 mapping calculation demonstrated short acquisition time and clearly high SNR. Results revealed that T1 relaxation time can be used as a biomarker to differentiate between healthy and pathologic Achilles tendon. However, T1 showed no correlation with the VISA-A clinical score.

scholarly journals Temporal concatenation for Markov decision processes

2021 ◽  
pp. 1-28
Author(s):  
Ruiyang Song ◽  
Kuang Xu
Keyword(s):  
Markov Decision Processes ◽  
Large Scale ◽  
Optimal Solution ◽  
Upper Bounds ◽  
Black Box ◽  
Decision Processes ◽  
Optimal Solutions ◽  
Wide Range ◽  
Markov Decision ◽  
Speed Up

We propose and analyze a temporal concatenation heuristic for solving large-scale finite-horizon Markov decision processes (MDP), which divides the MDP into smaller sub-problems along the time horizon and generates an overall solution by simply concatenating the optimal solutions from these sub-problems. As a “black box” architecture, temporal concatenation works with a wide range of existing MDP algorithms. Our main results characterize the regret of temporal concatenation compared to the optimal solution. We provide upper bounds for general MDP instances, as well as a family of MDP instances in which the upper bounds are shown to be tight. Together, our results demonstrate temporal concatenation's potential of substantial speed-up at the expense of some performance degradation.

Sign in / Sign up

Export Citation Format

Share Document