Optimal deployment of intelligent mobile air quality systems

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY MASTER’S GRADUATION THESIS Optimal deployment of intelligent mobile air quality systems NGUYEN VIET DUNG Dung.NV202342M@sis.hust.edu.vn Major: Data Science and Artificial Intelligence (Elitech) Thesis advisor: Institute: Assoc.Prof Do Phan Thuan _ School of Information and Communication Technology HA NOI, 09/2022 CỘNG HÒA XÃ HỘI CHỦ NGHĨA VIỆT NAM Độc lập – Tự – Hạnh phúc BẢN XÁC NHẬN CHỈNH SỬA LUẬN VĂN THẠC SĨ Họ tên tác giả luận văn: Nguyễn Việt Dũng Đề tài luận văn: Triển khai tối ưu hệ thống quan trắc khơng khí di động thơng minh Chun ngành: Khoa học liệu Trí tuệ nhân tạo Mã số SV: 20202342M Tác giả, Người hướng dẫn khoa học Hội đồng chấm luận văn xác nhận tác giả sửa chữa, bổ sung luận văn theo biên họp Hội đồng ngày 29/10/2022 với nội dung sau: - Thêm giới thiệu chi tiết nghiên cứu có liên quan chương - Đổi tên chương từ “Problem formulation & hardness” thành “Problem formulation” - Thêm phát biểu toán opportunistic sensing optimization trước viết tắt thành OSO - Đổi tên phần 3.2 thành “Mathematical formulation of OSO” - Thêm giải thích rõ hàm mục tiêu điều kiện mục 3.2 - Thêm lý giải thích sử dụng thuật toán quy hoạch động: “In this simplified scenario, our dynamic programming approach guarantees that the set found by the submaxSet function is always maximum thus the number 𝛼𝛼 mentioned in the previous section 5.1.1.2 will be equal to Later we will show that we cannot use dynamic programming in the general scenario, and we will need another greedy sub-process which has a lower performance ratio for that.” - Thêm số giải thích chi tiết thuật toán meta-heuristics lý lựa chọn sử dụng chúng, cụ thể sau: + “They are appropriate methods to verify efficiency of the approximation algorithm, since their tremendous performance in practice was shown in numerous research papers, especially researches related to air monitoring systems If the greedy approximation approach is decent, the experimental results produced SĐH.QT9.BM11 Ban hành lần ngày 11/11/2014 by it should be competitive to the ones produced by the chosen metaheuristics It is indeed true, and we will show the experimental results supporting this observation later in this thesis.” + “Two meta-heuristics, the genetic algorithm and the simulated annealing algorithm, are chosen to solve the OSO problem because of their simplicity and efficiency in practice Related researches about air monitoring systems also deployed these methods to solve challenging problems, and the results usually show that they are good choices for creating a solution.” - Thêm giải thích cho hình vẽ bảng biểu - Thêm mô tả input output cho thuật toán - Thêm mục 6.4 “Comparison of results between the approximation algorithm and the meta-heuristics” chuyển mục 6.4 cũ thành mục 6.5 “Discussion” Ngày Giáo viên hướng dẫn tháng năm Tác giả luận văn CHỦ TỊCH HỘI ĐỒNG SĐH.QT9.BM11 Ban hành lần ngày 11/11/2014 Graduation Thesis Assignment Name: Nguyen Viet Dung Phone: +84 399629097 Email : Dung.NV202342M@sis.hust.edu.vn Student ID: 20202342M Class: 20BKHDL-E Thesis title: Optimal deployment of intelligent mobile air quality systems Thesis code: 2020BKHDL-KH01 Affiliation : Hanoi University of Science and Technology I – Nguyen Viet Dung - hereby warrants that the work and presentation in this thesis performed by myself under the supervision of Assoc.Prof Do Phan Thuan All the results presented in this thesis are truthful and are not copied from any other works All references in this thesis including images, tables, figures and, quotes are clearly and fully documented in the bibliography I will take full responsibility for even one copy that violates school regulations Hanoi, 28th September, 2022 Author Nguyen Viet Dung Attestation of thesis advisor : I certify that the thesis entitled “Optimal deployment of intelligent mobile air quality systems” submitted for the degree of Master of Science (M.S.) by Mr Nguyen Viet Dung is the record of research work carried out by him during the period from 10/2020 to 10/2022 under my guidance and supervision, and that this work has not formed the basis for the award of any Degree, Diploma, Associateship and Fellowship or other Titles in this University or any other University or institution of Higher Learning Hanoi, 28th September, 2022 Thesis Advisor Assoc.Prof Do Phan Thuan Acknowledgements In order to obtain this master's thesis, apart from my own efforts, it is impossible not to mention the help of many other people First, I would like to thank Associate Professor Do Phan Thuan and Dr Nguyen Phi Le, my direct mentors From the time I got my thesis title to the time I finished it, there was not a moment that they didn't encourage me to run to the finish line I am where I am today in large part because of their support Next, I have to mention the funding source of VinIF Their financial support helped me to pay my tuition fees and complete my studies with peace of mind Finally, I would like to express my sincerest thanks to my teachers, friends and family Without them by my side, I wouldn't have made it to the end of the road Two years of wonderful lectures and extremely helpful time doing research will be in my heart forever Abstract Monitoring air quality plays a critical role in the sustainable development of developing regions where the air is severely polluted Air quality monitoring systems based on static monitors often not provide information about the area each monitor represents or represent only small areas In addition, they have high deployment costs that reflect the efforts needed to ensure sufficient quality of measurements Meanwhile, the mobile air quality monitoring system, such as the one in this work, shows the feasibility of solving those challenges The system includes environmental sensors mounted on buses that move along their routes, broadening the monitoring areas In such a system, we introduce a new optimization problem named opportunistic sensing that aims to find (1) optimal buses to place the sensors and (2) the optimal monitoring timing to maximize the number of monitored critical regions We investigate the optimization problem in two scenarios: simplified and general bus routes Initially, we mathematically formulate the targeted problem and prove its NP-hardness Then, we propose a polynomial-time -, 𝑒𝑒−1 2𝑒𝑒−1 - approximation algorithm for the problem with the simplified, general routes, respectively To show the proposed algorithms’ effectiveness, we have evaluated it on the real data of real bus routes in Hanoi, Vietnam The evaluation results show that the former algorithm guarantees an average performance ratio of 75.70%, while the latter algorithm achieves the ratio of 63.96% Notably, when the sensors can be on (e.g., enough energy) during the whole route, the 𝑒𝑒−1 2𝑒𝑒−1 -approximation algorithm achieves the approximation ratio of (1 − 𝑒𝑒) Such ratio, which is almost twice as 𝑒𝑒−1 , enlarges the average performance ratio to 78.42% 2𝑒𝑒−1 To further test the efficiency of the greedy approximation algorithm and optimize the results, we propose two more meta-heuristic algorithms for this problem: genetic algorithm and simulated annealing algorithm Experiments show that the above meta-heuristic algorithms only increase the goodness of the results by 1% to 3% on average, but have a much larger running time than the greedy algorithm From there, we see that the approximation algorithm in particular is already a feasible solution in practice without mentioning any other complicated tools Content Graduation Thesis Assignment Acknowledgements Abstract Content List of Figures List of Tables Acronyms 10 Chapter Introduction 11 1.1 Mobile air quality monitoring systems 11 1.2 Opportunistic sensing optimization (OSO) problem 12 1.3 Thesis contribution 12 1.4 Structure of thesis 12 Chapter Related works 13 Chapter Problem formulation 17 3.1 Problem statement 17 3.2 Mathematical formulation of OSO 18 3.3 Hardness of OSO 22 Chapter Theoretical background 24 4.1 Approximation algorithms 24 4.2 Meta-heuristic algorithms 24 4.3 Research methodology 27 Chapter Proposed solution 29 5.1 Approximation algorithms 29 5.2 Meta-heuristic algorithms 38 Chapter Experimental results 42 6.1 Experimental settings 42 6.2 Numerical results of approximation algorithms 45 6.3 Numerical results of meta-heuristic algorithms 51 6.4 Comparison of results between the approximation algorithm and the meta-heuristics 61 6.5 Discussion 61 Chapter Conclusion 63 Published papers 64 References 65 List of Figures Figure A map of size × with bus routes and critical squares When 𝑘𝑘 = 2, an example of the sensor’s turn-on positions on bus is shown With such selected positions, that sensor can observe critical squares 𝐴𝐴, 𝐵𝐵, 𝐶𝐶, 𝐷𝐷 and 𝐸𝐸 17 Figure Illustration of observable boundary, observable square, and observable segment 19 Figure Illustration of Theorem 3.1’s proof (𝑋𝑋 is an arbitrary point on a bus route segment 𝑃𝑃 𝑌𝑌 is the leftmost observable bound closest to 𝑋𝑋 If 𝐶𝐶 is a critical square observable by 𝑋𝑋, then it is also observable by 𝑌𝑌) .20 Figure A corresponding bus map when 𝛽𝛽 = 3, 𝑉𝑉 = {𝐴𝐴, 𝐵𝐵, 𝐶𝐶, 𝐷𝐷, 𝐹𝐹}, 𝑉𝑉 = {𝐴𝐴, 𝐶𝐶, 𝐷𝐷, 𝐸𝐸}, and 𝑉𝑉 = {𝐵𝐵, 𝐹𝐹} 23 Figure The remaining map after removing bus from the map in Fig 1, and the greedy process continues 29 Figure (a) [l Ab , 𝑟𝑟 Ab ] is the unique close segment that contains all sensor’s turn-on positions on the bus route 𝑏𝑏 where the critical square 𝐴𝐴 is observed (b) There are 𝑑𝑑 critical squares observed by turning on sensor from bus route 𝑏𝑏 (in this figure, 𝑑𝑑 = 5) Each square 𝑖𝑖 can be observed by a sensor turned on at somewhere in the middle of the interval [l ib , 𝑟𝑟 ib ] We then have 𝑑𝑑 critical points which are the left endpoints (l ib , where 𝑖𝑖 = 1, … , 𝑑𝑑) of such intervals 33 Figure Efficiency heatmap 45 Figure Performance in the simplified scenario with 𝑝𝑝 = 10, 𝑞𝑞 = 12 46 Figure Performance in the simplified scenario with 𝑝𝑝 = 25, 𝑞𝑞 = 30 47 Figure 10 Performance in the simplified scenario with 𝑝𝑝 = 30, 𝑞𝑞 = 36 47 Figure 11 Performance in the simplified scenario with 𝑝𝑝 = 42, 𝑞𝑞 = 50 47 Figure 12 Performance in the general and special scenario with 𝑝𝑝 = 10, 𝑞𝑞 = 12 48 Figure 14 Performance in the general and special scenario with 𝑝𝑝 = 30, 𝑞𝑞 = 36 49 Figure 15 Performance in the general and special scenario with 𝑝𝑝 = 42, 𝑞𝑞 = 50 50 List of Tables Table Notation list……………………………………………………………… …… 18 Table Meta-heuristics performance compared to the approximation algorithm’s results in the simplified scenario………………………………………………… ……………… 51 Table Meta-heuristics performance compared to the approximation algorithm’s results in the general scenario 55 25 x 30 10 25 50 100 200 1.0 63.16 0.031 63.16 0.00 1.848 63.16 0.00 69.516 2.0 74.00 0.059 78.00 +4.00 2.532 74.00 0.00 129.199 0.5 77.78 0.003 88.89 +11.11 2.845 88.89 +11.11 7.352 1.0 66.67 0.004 66.67 0.00 2.305 66.67 0.00 6.575 2.0 75.00 0.003 75.00 0.00 0.629 75.00 0.00 8.494 3.0 80.00 0.003 80.00 0.00 1.866 80.00 0.00 12.423 0.5 75.00 0.011 75.00 0.00 8.517 75.00 0.00 13.634 1.0 75.00 0.010 80.00 +5.00 7.825 80.00 +5.00 15.842 2.0 66.67 0.009 71.43 +4.76 4.053 71.43 +4.76 21.026 3.0 65.22 0.012 65.22 0.00 1.356 65.22 0.00 33.002 0.5 65.71 0.027 71.43 +5.72 4.281 71.43 +5.72 24.668 1.0 72.73 0.023 72.73 0.00 4.653 72.73 0.00 29.608 2.0 68.89 0.025 71.11 +2.22 5.944 68.89 0.00 44.425 3.0 61.91 0.029 61.91 0.00 1.837 61.91 0.00 71.817 0.5 62.65 0.072 67.47 +4.82 10.507 63.86 +1.20 49.002 1.0 58.21 0.068 58.21 0.00 3.677 58.21 0.00 59.469 2.0 62.64 0.097 61.54 -1.10 6.099 65.93 +3.30 93.112 3.0 62.50 0.123 62.50 0.00 3.280 62.50 137.105 0.5 60.82 0.239 64.33 +3.51 17.677 64.33 +3.51 104.522 1.0 53.10 0.285 53.10 0.00 9.853 53.10 0.00 0.00 151.528 56 300 400 30 x 36 10 25 50 2.0 57.38 0.387 58.47 +1.09 8.202 58.47 +1.09 214.758 3.0 62.91 0.746 62.25 -0.66 6.543 62.91 0.5 50.89 0.440 51.79 +0.89 11.955 51.79 +0.89 174.918 1.0 54.80 0.661 53.43 -1.37 15.331 54.80 2.0 55.99 1.272 58.45 +2.47 16.381 57.75 +1.76 351.679 3.0 62.89 2.565 62.89 0.00 5.235 0.5 50.34 0.858 50.34 0.00 18.649 52.04 +1.70 251.393 1.0 53.40 1.349 52.72 -0.68 17.310 53.40 2.0 56.76 3.080 57.57 +0.81 33.678 57.84 +1.08 573.461 3.0 63.07 6.886 64.46 +1.39 23.599 63.07 0.00 1389.834 0.5 77.78 0.003 77.78 0.00 4.151 77.78 0.00 1.0 77.78 0.003 88.89 +11.11 4.908 88.89 +11.11 7.201 2.0 75.00 0.003 75.00 0.00 0.755 75.00 0.00 10.635 3.0 83.33 0.003 83.33 0.00 0.826 83.33 0.00 11.361 0.5 84.21 0.010 84.21 0.00 6.806 84.21 0.00 12.698 1.0 60.00 0.010 60.00 0.00 5.489 60.00 0.00 14.324 2.0 76.00 0.009 80.00 +4.00 5.559 76.00 0.00 19.747 3.0 70.83 0.009 70.83 0.00 4.603 70.83 0.00 23.893 0.5 66.67 0.027 66.67 0.00 6.502 64.10 -2.56 23.598 1.0 62.86 0.025 62.86 0.00 3.180 62.86 31.891 62.89 0.00 0.00 0.00 0.00 0.00 347.961 249.174 797.241 391.696 6.874 57 100 200 300 400 500 2.0 58.54 0.025 68.29 +9.76 4.499 68.29 +9.76 37.372 3.0 62.16 0.029 62.16 0.00 2.384 62.16 0.00 58.677 0.5 69.44 0.066 70.83 +1.39 9.913 70.83 +1.39 42.094 1.0 63.64 0.067 67.53 +3.90 35.936 66.23 +2.60 46.238 2.0 55.56 0.082 55.56 0.00 6.782 56.67 +1.11 78.931 3.0 62.69 0.104 62.69 0.00 6.084 62.69 0.00 131.681 0.5 66.91 0.205 66.91 0.00 16.704 66.91 0.00 87.527 1.0 54.17 0.250 54.17 0.00 12.160 54.17 0.00 115.716 2.0 54.95 0.354 56.04 +1.10 10.100 56.04 +1.10 189.385 3.0 62.90 0.529 62.90 0.00 7.966 62.90 0.00 287.857 0.5 60.77 0.442 60.29 -0.48 20.091 60.77 0.00 143.786 1.0 52.34 0.568 55.32 +2.98 19.337 55.32 +2.98 187.807 2.0 55.20 1.095 53.76 -1.43 17.690 55.56 +0.36 325.873 3.0 61.81 1.969 66.32 +4.51 21.368 65.28 +3.47 431.393 0.5 52.30 0.769 51.97 -0.33 21.101 52.63 +0.33 197.473 1.0 50.14 1.070 50.14 0.00 25.388 50.14 2.0 52.98 1.993 53.26 +0.28 20.407 53.26 +0.28 479.796 3.0 62.96 4.006 62.96 0.00 20.036 62.96 0.00 1039.680 0.5 51.50 1.140 51.50 0.00 28.696 51.50 0.00 1.0 53.72 1.751 54.65 +0.93 51.960 54.19 +0.47 370.321 0.00 282.635 283.167 58 42 x 50 10 25 50 100 200 2.0 53.44 3.777 54.32 +0.89 60.282 54.10 +0.66 622.763 3.0 60.55 10.806 61.83 +1.28 42.663 61.19 +0.64 754.858 0.5 100.00 0.003 100.00 0.00 0.602 100.00 0.00 4.324 1.0 88.89 0.003 88.89 0.00 3.451 88.89 0.00 6.030 2.0 70.00 0.003 70.00 0.00 1.564 70.00 0.00 8.257 3.0 62.50 0.003 62.50 0.00 1.355 62.50 0.00 8.946 0.5 83.33 0.011 83.33 0.00 1.349 83.33 0.00 11.212 1.0 77.78 0.010 77.78 0.00 3.878 77.78 0.00 12.074 2.0 79.17 0.010 79.17 0.00 8.963 75.00 -4.17 15.365 3.0 63.64 0.010 72.73 +9.09 1.077 63.64 0.00 19.225 0.5 65.63 0.027 65.63 0.00 6.454 65.63 0.00 19.384 1.0 74.36 0.028 79.49 +5.13 8.944 76.92 +2.56 21.638 2.0 62.50 0.026 65.00 +2.50 4.026 65.00 +2.50 30.237 3.0 59.09 0.027 63.64 +4.55 7.275 63.64 +4.55 36.256 0.5 70.31 0.071 73.44 +3.13 13.921 70.31 1.0 65.82 0.066 67.09 +1.27 2.0 59.14 0.078 62.37 3.0 56.32 0.086 0.5 69.57 1.0 61.33 0.00 35.943 8.162 67.09 +1.27 42.741 +3.23 6.193 62.37 +3.23 62.031 56.32 0.00 3.733 56.32 0.00 81.768 0.200 72.46 +2.90 32.199 71.74 +2.17 69.035 0.198 61.33 0.00 13.926 61.33 89.575 0.00 59 300 400 500 1000 2.0 54.48 0.256 53.79 -0.69 6.271 3.0 53.89 0.340 57.78 +3.89 11.633 57.78 +3.89 179.273 0.5 62.30 0.391 63.39 +1.09 25.115 62.30 1.0 54.74 0.426 55.26 +0.53 12.106 55.26 +0.53 145.949 2.0 55.35 0.640 52.03 -3.32 20.631 55.35 3.0 51.64 0.842 54.55 +2.91 13.443 54.55 +2.91 295.075 0.5 56.03 0.584 54.86 -1.17 19.659 56.03 0.00 158.071 1.0 53.46 0.664 53.46 0.00 16.624 53.46 0.00 214.609 2.0 54.93 1.128 54.93 0.00 17.387 54.93 0.00 343.229 3.0 51.37 2.096 51.10 -0.28 25.342 52.75 +1.37 438.352 0.5 62.58 0.999 61.98 -0.60 34.305 62.58 0.00 196.484 1.0 55.56 1.196 55.56 0.00 30.234 55.56 0.00 252.401 2.0 57.44 2.148 57.44 0.00 22.872 57.44 0.00 494.512 3.0 50.97 4.239 51.18 +0.21 32.244 52.26 +1.29 636.920 0.5 49.26 4.447 49.26 0.00 60.543 49.26 1.0 52.17 7.101 52.32 +0.15 72.606 52.32 +0.15 801.631 2.0 54.83 16.733 56.05 +1.22 191.155 55.49 +0.67 1063.719 3.0 53.51 40.125 53.51 0.00 Avg efficiency change (%) +1.28 54.48 64.266 53.51 0.00 0.00 0.00 0.00 140.323 106.710 201.175 534.257 0.00 2192.211 +0.91 60 6.4 Comparison of results between algorithm and the meta-heuristics the approximation Based on the tables in section 6.3, we can see that meta-heuristics outperformed the approximation algorithm by just a small margin (from 1% to 3% on average), though consuming much more running time On average, GA helps increase the minimum efficiency by 2.69% in the simplified case and 1.28% in the general case, while SA helps increase the minimum efficiency by 1.94% in the simplified case and 0.91% in the general case In spite of that, there is still a considerable number of times where meta-heuristics are better, which is quite promising Out of all 172 test cases including simplified and general scenarios, there are 86 times (50% of time) that GA is better than the approximation algorithm, and for SA there are 72 times (41.86% of time) There are only 16 times (9.3% of time) that GA is worse than the approximation algorithm, and for SA there are only 13 times (7.56% of time) These numbers mean that meta-heuristics are usually a better choice for efficiency GA is slightly better than SA in this problem, however the difference is pretty small, and there are still cases where SA outperformed GA To be practical, an algorithm should run fast enough We see that meta-heuristics improves the quality of the results, but their running time is a big problem Compared to the approximation approach, GA and SA are usually hundred or even thousand times slower, but does not guarantee a better solution In these meta-heuristics, the bigger the hyperparameters, the more chance that a good result will be found However making those hyperparameters bigger than the current setup is not a suitable act since the current running time of our meta-heuristics are already huge, and it would take days or months to try a new set of hyperparameters, which is infeasible All of these observations suggest that our approximation algorithm’s results were not far from optimal However, we should still consider the meta-heuristics approaches in real life if time budget allows 6.5 Discussion This research is in line with the implementation of the mobile air quality monitoring system in Hanoi, as presented in [24] With the system, the numerical results obtained from this study can be applied in our real system as follows As the Hanoi bus map has the size of 20 km × 30 km, and the sensing range of the used sensor is about 120 m, our network topology corresponds to the test case of 𝑝𝑝 = 42, 𝑞𝑞 = 50, and 𝑟𝑟 = 0.5 Figs 10, 14 depict the minimum efficiency and running time obtained in this test case As shown, for all combinations of (𝑚𝑚, 61 𝑘𝑘) (i.e., the number of sensors and the number of turn-on times of each sensor), our proposed solution achieved the minimum efficiency from 49.26% to 100% when the number of critical regions varies from 10 to 1000 Moreover, when the number of critical regions is 1000 and the sensors are sufficiently powered by the bus to be turned on all the time during the bus routes, the number of sensors required to cover the maximum number of critical regions is 38 When the number of sensors is reduced to 16, we achieve 90% of the maximal coverage We also deduce that to save about 95 percent of sensors’ energy consumption, we need about 40 sensors This study presented a theoretical approach to address the opportunistic sensing in mobile air quality monitoring systems We have made some ideal assumptions, including the diskbased sensing paradigm and the unchanging nature of the critical regions In practice, the sensing model is more complicated, and the measurement accuracy may be affected by many factors such as temperature, humidity [25] Besides, the critical regions can also change over time For such additional problems, we need to leverage other techniques (e.g., machine learning [26, 27]) that can capture the network’s dynamic nature and forecast its future state; thereby, allowing for decision making that is adaptable to the network’s variety These problems will be addressed in our future work 62 Chapter Conclusion This thesis investigated the opportunistic sensing optimization (OSO) problem in the mobile air quality monitoring system First, we have mathematically formulated the OSO problem and proved its NP-hardness Second, we proposed a polynomial-time 𝑒𝑒−1 2𝑒𝑒−1 − and −approximation algorithms, which are for OSO in the general scenario and the simplified one, respectively Finally, we proposed two meta-heuristics approaches: genetic algorithm and simulated annealing, to further test the efficiency of the greedy approximation algorithm and optimize the results We have extensively evaluated the approximation algorithms on the real bus map, similar to the one in Hanoi, Vietnam The evaluation results showed that the proposed algorithm’s average performance ratio is 63.96% for the general scenario and 75.70% for the simplified one Moreover, the maximum value of average running time is 40.125 seconds among all test cases Especially, if the energy supply is sufficient for the sensor to be turned on for the whole route (i.e., the special case where 𝑘𝑘 ≥ 𝑐𝑐), the approximation ratio will become (1 − 𝑒𝑒), which is almost twice as much as 78.42% 𝑒𝑒−1 , and also the average performance ratio becomes 2𝑒𝑒−1 For the meta-heuristic methods, experiments show that these algorithms only increase the goodness of the results by 1% to 3% on average, but have a larger running time than the greedy algorithm From there, we see that the approximation algorithm in particular is already a feasible solution in practice without mentioning any other complicated tools However, when time budget allows, meta-heuristics should be considered since they usually produce better solutions 63 Published papers [1] Nguyen, V D., Le Nguyen, P., Nguyen, K., & Do, P T (2022) Constant approximation for opportunistic sensing in mobile air quality monitoring system Computer Networks, 202, 108646 [2] V.D Nguyen, P Le Nguyen, T.H Nguyen, K Nguyen, P.T Do, An 2𝑒𝑒−1 - 𝑒𝑒−1 Approximation Algorithm for Maximizing Coverage Capability in Mobile Air Quality Monitoring Systems, in: Proc IEEE NCA, 2020, pp 1–4 [3] V.D Nguyen, P Le Nguyen, T.H Nguyen, P.T Do, A -Approximation Algorithm for Target Coverage Problem in Mobile Air Quality Monitoring Systems, in: Proc IEEE GLOBECOM, 2020, pp 1–6 64 References [1] Ambient air pollution: A global assessment of exposure and burden of disease, 2021, (Access date: 2021 June), https://www.who.int/phe/publications/air-pollution-globalassessment/en/ [2] A Lozano, J Usero, E Vanderlinden, J Raez, J Contreras, B Navarrete, Air quality monitoring network design to control nitrogen dioxide and ozone, applied in Malaga, Spain, Microchem J 93 (2) (2009) 164–172 [3] O Postolache, M Pereira, P Girao, Smart Sensor Network for Air Quality Monitoring Applications, in: Proc IEEE Instrumentation and Measurement Technology Conference, 2005, pp 537–542 [4] J.-H Liu, Y.-F Chen, T.-S Lin, D.-W Lai, T.-H Wen, C.-H Sun, J.-Y Juang, J.Jiang, Developed urban air quality monitoring system based on wireless sensor networks, in: Proc IEEE ICST, 2011, pp 549–554 [5] K Zheng, S Zhao, Z Yang, X Xiong, W Xiang, Design and implementation of LPWA-based air quality monitoring system, IEEE Access (2016) 3238–3245 [6] R Yasmin, J Petäjäjärvi, K Mikhaylov, A Pouttu, Large and Dense LoRaWAN Deployment to Monitor Real Estate Conditions and Utilization Rate, in: Proc IEEE PIMRC, 2018, pp 1–6 [7] Air visual, 2021, (Access date: 2021 June), https://www.airvisual.com/vietnam/hanoi [8] T Villa, F Gonzalez, B Miljievic, Z Ristovski, Morawska, An overview of small unmanned aerial vehicles for air quality measurements: Present applications and future prospectives, Sensors 16 (2016) [9] Y Yang, Z Bai, Z Hu, Z Zheng, K Bian, L Song, AQNet: Fine-grained 3D spatiotemporal air quality monitoring by aerial-ground WSN, in: Proc IEEE INFOCOM WKSHPS, 2018, pp 1–2 [10] S Kaivonen, E.C.-H Ngai, Real-time air pollution monitoring with sensors on city bus, Digit Commun Netw (1) (2020) 23–30 65 [11] S.M Biondi, V Catania, S Monteleone, C Polito, Bus as a sensor: A mobile sensor nodes network for the air quality monitoring, in: Proc IEEE WiMob, 2017, pp 272– 277 [12] X Cheng, D.-Z Du, L Wang, B Xu, Relay sensor placement in wireless sensor networks, Wirel Netw 14 (3) (2008) 347–355 [13] F Senel, M Younis, Relay node placement in structurally damaged wireless sensor networks via triangular steiner tree approximation, Comput Commun 34 (16) (2011) 1932–1941 [14] A Shan, X Xu, Z Cheng, Target coverage in wireless sensor networks with probabilistic sensors, Sensors 16 (9) (2016) [15] X Zhu, J Li, M Zhou, Target coverage-oriented deployment of rechargeable directional sensor networks with a mobile charger, IEEE Internet Things J (3) (2019) 5196–5208 [16] D Arivudainambi, S Balaji, T.S Poorani, Sensor deployment for target coverage in underwater wireless sensor network, in: Proc International Conference on Performance Evaluation and Modeling in Wired and Wireless Networks (PEMWN), 2017, pp 1–6 [17] N.T Hanh, H.T.T Binh, N Van Son, P.N Lan, Minimal Node Placement for Ensuring Target Coverage With Network Connectivity and Fault Tolerance Constraints in Wireless Sensor Networks, in: Proc IEEE CEC, 2019, pp 2923–2930 [18] X Gao, Z Chen, F Wu, G Chen, Energy efficient algorithms for 𝑘𝑘-sink minimum movement target coverage problem in mobile sensor network, IEEE/ACM Trans Netw 25 (6) (2017) 3616–3627 [19] N.T Nguyen, B Liu, S Wang, On new approaches of maximum weighted target coverage and sensor connectivity: Hardness and approximation, IEEE Trans Netw Sci Eng (2019) [20] M Rout, R Roy, Self-deployment of mobile sensors to achieve target coverage in the presence of obstacles, IEEE Sens J 16 (14) (2016) 5837–5842 [21] R Choudhuri, R.K Das, Coverage of targets in mobile sensor networks with restricted mobility, IEEE Access (2018) 10803–10813 66 [22] D.S Hochbaum, Approximating covering and packing problems: set cover, vertex cover, independent set, and related problems, in: Approximation Algorithms for NPHard Problems, PWS Publishing Company, 1996, pp 94–143 [23] C Chekuri, A Kumar, Maximum coverage problem with group budget constraints and applications, in: Approximation, Randomization, & Combinatorial Optimization: Algorithms & Techniques, 2004, pp 72–83 [24] V.A Nguyen, V.H Vu, V.S Doan, T.H Nguyen, P.T Do, K Nguyen, P.L Nguyen, M.T Le, Realizing mobile air quality monitoring system: Architectural concept and device prototype, in: Asia Pacific Conference on Communications (APCC), 2021 [25] M Balanescu, I Oprea, G Suciu, M.-A Dobrea, C Balaceanu, R.-I Ciobanu, C Dobre, A study on data accuracy for IoT measurements of PMs concentration, in: 2019 22nd International Conference on Control Systems and Computer Science (CSCS), 2019, pp 182–187 [26] M Dobrea, A Bădicu, M Barbu, O Subea, M Bălănescu, G Suciu, A Bỵrdici, O Orza, C Dobre, Machine learning algorithms for air pollutants forecasting, in: 2020 IEEE 26th International Symposium for Design and Technology in Electronic Packaging (SIITME), 2020, pp 109–113 [27] M.H Nguyen, P Le Nguyen, K Nguyen, V.A Le, T.-H Nguyen, Y Ji, PM2.5 prediction using genetic algorithm-based feature selection and encoder-decoder model, IEEE Access (2021) 57338–57350 [28] Michalewicz Z, Schoenauer M (1996), Evolutionary algorithms for constrained parameter optimization problems Evol Comput 4(1):1–32 [29] Holland JH (1975), Adaptation in natural and artificial systems The U of Michigan Press [30] Scott Kirkpatrick, C Daniel Gelatt, and Mario P Vecchi Optimization by simulated annealing science, 220(4598):671–680, 1983 https://doi.org/10.1126/science.220.4598.671 [31] Vladimír Černỳ Thermodynamical approach to the traveling salesman problem: An efficient simulation algorithm Journal of optimization theory and applications, 45(1):41–51, 1985 https://doi.org/10.1007/bf00940812 [32] El-Ghazali Talbi Metaheuristics: from design to implementation, volume 74 John Wiley & Sons, 2009 https://doi.org/10.1002/9780470496916 67 CỘNG HÒA XÃ HỘI CHỦ NGHĨA VIỆT NAM Độc lập – Tự – Hạnh phúc BẢN XÁC NHẬN CHỈNH SỬA LUẬN VĂN THẠC SĨ Họ tên tác giả luận văn: Nguyễn Việt Dũng Đề tài luận văn: Triển khai tối ưu hệ thống quan trắc không khí di động thơng minh Chun ngành: Khoa học liệu Trí tuệ nhân tạo Mã số SV: 20202342M Tác giả, Người hướng dẫn khoa học Hội đồng chấm luận văn xác nhận tác giả sửa chữa, bổ sung luận văn theo biên họp Hội đồng ngày 29/10/2022 với nội dung sau: - Thêm giới thiệu chi tiết nghiên cứu có liên quan chương - Đổi tên chương từ “Problem formulation & hardness” thành “Problem formulation” - Thêm phát biểu toán opportunistic sensing optimization trước viết tắt thành OSO - Đổi tên phần 3.2 thành “Mathematical formulation of OSO” - Thêm giải thích rõ hàm mục tiêu điều kiện mục 3.2 - Thêm lý giải thích sử dụng thuật toán quy hoạch động: “In this simplified scenario, our dynamic programming approach guarantees that the set found by the submaxSet function is always maximum thus the number 𝛼𝛼 mentioned in the previous section 5.1.1.2 will be equal to Later we will show that we cannot use dynamic programming in the general scenario, and we will need another greedy sub-process which has a lower performance ratio for that.” - Thêm số giải thích chi tiết thuật toán meta-heuristics lý lựa chọn sử dụng chúng, cụ thể sau: + “They are appropriate methods to verify efficiency of the approximation algorithm, since their tremendous performance in practice was shown in numerous research papers, especially researches related to air monitoring systems If the greedy approximation approach is decent, the experimental results produced by it should be competitive to the ones produced by the chosen metaheuristics It is indeed true, and we will show the experimental results supporting this observation later in this thesis.” + “Two meta-heuristics, the genetic algorithm and the simulated annealing algorithm, are chosen to solve the OSO problem because of their simplicity and efficiency in practice Related researches about air monitoring systems also deployed these methods to solve challenging problems, and the results usually show that they are good choices for creating a solution.” - Thêm giải thích cho hình vẽ bảng biểu - Thêm mô tả input output cho thuật toán - Thêm mục 6.4 “Comparison of results between the approximation algorithm and the meta-heuristics” chuyển mục 6.4 cũ thành mục 6.5 “Discussion” Ngày Giáo viên hướng dẫn tháng Tác giả luận văn CHỦ TỊCH HỘI ĐỒNG năm ... Viet Dung Attestation of thesis advisor : I certify that the thesis entitled ? ?Optimal deployment of intelligent mobile air quality systems” submitted for the degree of Master of Science (M.S.) by... 20202342M Class: 20BKHDL-E Thesis title: Optimal deployment of intelligent mobile air quality systems Thesis code: 2020BKHDL-KH01 Affiliation : Hanoi University of Science and Technology I – Nguyen... Number of columns in the bus map grid q Number of rows in the bus map grid c Number of critical squares n Number of bus routes m Number of sensors r Sensing radius of a sensor k Maximum number of