https://doi.org/10.31449/inf.v48i9.5903 Informatica 48 (2024) 113–126 113 Effect Analysis of Adaptive Ant Colony Algorithm with QoS Constraints Applied in Drone Disaster Area Search and Rescue System Yanyan Dou School of Information Engineering, Zhengzhou University of Technology, Zhengzhou, 450000, China E-mail: douyanyanddd@163.com Keywords: drones, disaster area search and rescue, service quality, ant colony, traveler length Received: March 14, 2024 With the development of technology, drones have been widely used in disaster area search and rescue. This study proposes an adaptive ant colony algorithm based on service quality constraints for drone search and rescue models. The algorithm aims to address the low efficiency and long-time consumption in drone disaster area search and rescue. The model first constructs a disaster area drone search and rescue scenario, then introduces service quality constraints to optimize and improve the ant colony algorithm. Finally, Matlab R2017a is used for simulation experiments to test the algorithm’ s performance. The experiment conducted in a 32-scale area showed that the proposed algorithm achieved a minimum travel dealer length of 1.55km, which is 0.05km and 0.01km less than the traditional ant colony algorithm’ s 1.60km and particle swarm optimization algorithm’ s 1.56km, respectively. In the 128 scale and 512 scale regions, the proposed algorithm exhibited shorter average path lengths and higher stability. The data shows that research algorithms have better search and rescue efficiency and quality in dynamically changing disaster environments. Povzetek: Prilagodljiv algoritem mravljiščne kolonije z omejitvami kakovosti storitve izboljšuje učinkovitost in stabilnost brezpilotnih letal pri iskanju in reševanju v območjih nesreč, kar zmanjša povprečno dolžino poti. 1 Introduction With the increasing frequency of natural disasters, disaster response and disaster Search and Rescue (S/R) works have received great attention. Against the backdrop of difficult S/R efforts in disaster areas, drone technology has emerged in people's sight [1]. Drones have become an important tool in Disaster Area Search and Rescue (DAS/R) due to their high flexibility, convenience, and efficiency [2]. Through drone technology, disaster relief can conduct aerial surveys of the disaster area, locate trapped individuals, evaluate the level of disaster, and effectively improve the efficiency of disaster relief [3]. However, the sudden and strong disaster environment poses higher requirements for the path planning of unmanned aerial vehicles. At present, the number of drones in disaster relief is constantly increasing, but how to set the optimal S/R route for drones is still a problem that needs to be solved [4]. In Path Planning Algorithms (PPA), Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) algorithms are common and have certain effectiveness in static scenes. However, in the face of dynamic and changing disaster environments, there are significant shortcomings in the stability and adaptability of ACO and PSO. Therefore, to optimize the S/R path of drones in disaster areas, improve the efficiency and quality of S/R tasks, this study proposes a Quality of Service Constrained Adaptive Ant Colony Optimization Algorithm (QSACO) based on service quality constraints. This study constructs a simulation environment for complex disaster areas, comprehensively considers the constraints of pheromone mechanism and Quality of Service (QoS), and improves its adaptability and efficiency in dynamic environments by optimizing algorithm parameters. The innovation of the research method is to constrain QoS to ACO and adaptively optimize the parameters in the ACO algorithm to improve the path planning efficiency of drones in DAS/R. The research content consists of four parts. Part 1 summarizes the research achievements of domestic and foreign scholars on drone S/R systems and PPA, analyzes the shortcomings of current research, and proposes the advantages of research methods. Part 2 is to build a disaster area environment, analyze the drone S/R process, and introduce QoS constraints into the ACO algorithm to build a model about the QSACO algorithm. Part 3 conducts performance analysis on the constructed QSACO algorithm, adjusts key parameters of the algorithm through simulation, and analyzes the impact of parameters on S/R results. Part 4 summarizes the experimental results and analyzes the shortcomings in the 114 Informatica 48 (2024) 113–126 Y. Dou research, proposing future research directions. 2 Related works When facing natural disasters, conducting S/R missions in disaster areas is urgent and important. Therefore, many scholars have conducted research on the DAS/R system of unmanned aerial vehicles. Albanese et al. proposed a S/R solution based on drones, which utilizes the high penetration rate of mobile phones in society to locate missing persons. Through on-site restoration experiments, it had been proven that this method could quickly locate missing victims through mobile phones, and had the characteristic of low battery consumption cost [5]. Dong et al. collected a dataset of thermal images captured by drones and conducted experimental analysis using different types of deep convolutional networks. The optimal point for pruning and fine-tuning the survivor detection network based on convolutional layer sensitivity was validated on NVIDIA's Jetson TX2 and achieved real-time performance of 26.60 frames per second [6]. Ribeiro et al. developed a mixed integer linear programming model for synchronous networks. This model considered the use of drones and mobile charging stations in a synchronous manner and was improved through genetic algorithms to adjust heuristic methods. This model had been applied in practical S/R missions, indicating that it provides efficient and fast response in S/R tasks [7]. Atif et al. proposed a customized online graph traversal method. This method minimized the search energy of drones and the number of unoccupied nodes, and further optimized the flight path and search energy of drones in the probability region through clustering. The proposed method could achieve linear scaling of the search area relative to the noise level. For a given noise level and an increasing number of nodes, this method had the drones search energy of the cluster, which can reduce the energy cost to 70% [8]. The core of drone’s DAS/R is PPA. In terms of PPA, there are already many and mature studies. Li et al. proposed a drone PPA based on an improved ACO approach. The algorithm utilized a grid method to create a 3D model of the target planning area for the drones. Secondly, the updating rules of pheromones were improved, and the weight factors of pheromones and heuristics were adjusted. The improved algorithm had better flight paths than traditional algorithms [9]. Mandloi et al. proposed a comparative analysis of path planning strategies for A-star, Theta-star, and Lazy Theta-star algorithms in a 3D environment. It verified the algorithm by calculating two performance indicators: time and path length. A-star guaranteed to find the shortest path on the graph, but could not guarantee to find it in a real continuous environment. The Lazy Theta-star algorithm could overcome the shortcomings of the A-star algorithm, but it would increase computational time [10]. Gul et al. proposed a multi-objective PPA, which first optimizes the path using the grey wolf optimizer PSO hybridization, and then combines all optimal and feasible points with local search techniques. Finally, relying on collision avoidance and detection algorithms, the mobile robot detected obstacles in its perception circle. This study conducted different simulations in various environments and found that the algorithm produces more feasible paths over shorter distances [11]. Chi et al. proposed a heuristic PPA based on a generalized Voronoi graph to generate heuristic paths, guide the sampling process of RRTs, and further improve the motion planning efficiency of RRTs. A series of simulations and practical verification have confirmed that the proposed algorithm has achieved better performance in heuristic path planning, free space feature extraction, and real-time motion planning [12]. In summary, the current PPA performs well in a preset environment. However, in a dynamically changing disaster area environment, the algorithm's performance may be challenged due to a lack of real-time response to changes and path optimization. The specific advantages and disadvantages of the above methods are shown in Table 1. Therefore, in view of the problems in Table 1, QSACO algorithm is proposed and QoS constraints are expected to be introduced to adapt to the changes in the disaster area environment. Table 1: Summary of related work Method Author Property Advantage Shortcoming Mobile phone-based drone search Albanese et al. [5] Use mobile phone location, fast response, suitable for densely populated areas; Low energy design, extended operating time. Relying on mobile phone signal coverage, the effect is limited when the signal is poor in the disaster area; May be limited due to privacy concerns. Thermal image and deep convolutiona l networks Dong et al. [6] High search accuracy; Strong real-time processing capability. High requirements on computing resources require dedicated hardware support. The recognition rate may be reduced by environmental influences, such as smoke and dust. Effect Analysis of Adaptive Ant Colony Algorithm with QoS… Informatica 48 (2024) 113–126 115 Mixed integer linear programmin g model Ribeiro et al. [7] Combined with the use of drones and mobile charging stations, it increases the continuity and flexibility of operations. Provides a fast and efficient scheduling policy. The solution of the model is complex and requires a long calculation time. In practical applications, frequent adjustments may be required due to large environmental changes. Traversal method in line graph Atif et al. [8] Significantly reduce search energy consumption and improve S/R efficiency. The search path is optimized by clustering to reduce the invalid coverage area. The initial knowledge and node distribution of the region are required. When environmental information is incomplete, the effect may decrease. 3 Drones DAS/R system with adaptive ACO based on QoS constraints For the drone’s DAS/R system, this study constructs a S/R scenario using drones in disaster areas, analyzes the S/R process, and optimizes drone’s S/R using ACO. In the application process of ACO, QoS is introduced to improve and optimize the efficiency and quality of drones S/R. 3.1 Construction of Drones S/R scenes in disaster areas The actual DAS/R involves a variety of scenario conditions, including Disaster Stricken Base Stations (DSBS), drones, and other devices. DSBS is used to build connections between users and Geographic Information System (GIS) databases, including Central Base Stations (CBS) and Temporary Base Stations (TBS) [13]. The setting of the CBS needs to cover all TBSs so that they can upload priority level information. Figure 1 shows the positional relationship between the two. Base station center Temporary base station Unmanned Aerial Vehicle Figure 1: Position relationship between CBS and TBS In Figure 1, CBS is located between each TBS. CBS obtains various geographic information indicators of 116 Informatica 48 (2024) 113–126 Y. Dou disaster-stricken areas by applying for GIS access permissions. At the same time, it combines the severity of the disaster area transmitted by TBS to the main base station, as well as the level of demand for disaster relief materials, to obtain normalized processing of various data. Furthermore, the S/R priority in the disaster area is obtained, and the initialization parameters of the PPA are determined. Drones used for transporting disaster relief supplies and finding routes are dispatched from the base station to traverse all sub-disaster areas near the highest priority affected area. Until it is fully traversed, drones return to the base station and wait for the next round of S/R commands [14]. In the DAS/R scenario, Figure 2 shows the preliminary preparation work for drones. Base station center Temporary base station Collect information and return it Information processing and assigning different priorities to disaster-affected areas Disaster area topology with weights Disaster area UAV Figure 2: Scene construction of DAS/R In Figure 2, the drones group is centrally managed by the base station. The main base station instructed a group of drones to perform a breadth-first search of the disaster area using geographic location information obtained from GIS. The drones group starts from the base station and prioritizes reaching the sub disaster areas adjacent to the base station until all sub disaster areas in the entire disaster area are traversed. Considering the dynamic characteristics of the disaster scene, new disaster areas may appear at any time. This study uses the new path exploration property of ACO to explore new disaster areas that GIS did not record in a timely manner, while prioritizing the breadth of drones search. 3.2 Analysis of ACO's solution to disaster relief issues ACO is a path search algorithm based on ant foraging behavior. During the process of foraging, ants leave behind an information substance on their path, which is called a pheromone. Pheromones can guide later ants to choose paths with higher pheromone concentrations for foraging. Due to the large number of ants passing through a short path in a unit of time, the short path will leave behind a higher concentration of pheromones. Subsequent ants are likely to follow this path for foraging due to the stronger pheromones present. As the pheromone trail becomes stronger, more ants are expected to follow this path. This is the phenomenon of positive feedback of information [15]. A B C D E F d=1 d=1 d=0.5 d=0.5 Ant*50 A B C D E F Ant*50 Ant*25 Ant*25 Ant*25 Ant*25 A B C D E F Ant*50 Ant*20 Ant*20 Ant*30 Ant*30 Figure 3: Schematic diagram of ACO path In Figure 3, A represents the ant nest, and CD represents the obstacles encountered by ants during their foraging process. Obstacles divide the path into two parts, one is path BCE and the other is path BDF. F represents the target food, and d represents the path distance. This study assumes that 50 ants travel from point A to point F, and Effect Analysis of Adaptive Ant Colony Algorithm with QoS… Informatica 48 (2024) 113–126 117 when faced with path selection, there are no pheromones present on the initial path. Therefore, ants will randomly choose a path with equal probability. Over time, 30 ants passed through BDE, while only 20 ants covered the path on BCE. This means that the concentration of pheromones left on the BDE pathway is 1.5 times higher than that on the BCE pathway. If 50 more ants depart from the ant nest or food source, these 30 ants will be influenced by the pheromones left by the previous ants when facing path selection, with 30 choosing BDE and only 20 choosing BCE [16]. In this research scenario, disaster relief is essentially a traveling salesman problem, so using ACO to solve disaster relief problems has significant advantages. Ants select their next path during foraging based on the concentration of residual substances in the path. This study uses k tabu to represent the current disaster area that ants have passed through. Taboo table k tabu will be updated with the dynamic behavior of ants. During the journey, ants will calculate the corresponding Path Transition Probability (PTP) based on the content of pheromones and heuristic information on each path. The specific calculation is formula (1) [17]. [ ( )] [ ( )] [ ( )] [ ( )] 0 k ij ik k ij ik ij s allowed tt tt p         =     (1) In formula (1), p represents the probability of path transition. k allowed represents the disaster area that has not yet been passed through.  represents the heuristic factor of information, mainly reflecting the impact of accumulated information on subsequent ants along the path. The magnitude of its value determines the strength of the collaborative ability within the ant colony. This parameter emphasizes the role of heuristic information in ants' decision-making process. A higher value of  makes ants more inclined to choose the path with higher heuristic value, which promotes the search of problem-specific heuristic information.  indicates the importance of heuristic information to the path selection process of ants, which directly affects the degree of dependence of ants on heuristic information in the path selection process. () ij t  represents the heuristic function, and its expression is formula (2). 1 () ij ij t d  = (2) In formula (2), d represents the distance between disaster areas. The smaller the value of d and the larger the value of () ij t  , the higher the probability of the path being selected. The update method of ant pheromones is formula (3). 1 ( ) (1 ) ( ) ( ) ( ) ( ) ij ij ij m k ij ij k t n t t tt      = + = −  +      =     (3) In formula (3),  represents the volatility coefficient of pheromones, with a range of values between [0,1) to avoid continuous accumulation of pheromones. 1  − represents the degree of residual pheromones. The evaporation coefficient of pheromone can be used to simulate the natural phenomenon of pheromone disappearing with time in real ant colonies. A higher volatilization coefficient indicates that pheromones dissipate more quickly. This helps the algorithm avoid premature convergence to local optimal solutions and improves its ability to explore new paths. () ij t   represents the amount of pheromone increase on the path. This defines the concentration increase of pheromones on the path after the ant has walked it. This mimics ant leaving pheromones when they return to the nest after finding a food source. The pheromones guide later ants to the food source. k ij   represents the residual content of pheromones along the path. According to the above steps, the ACO process is Figure 4. 118 Informatica 48 (2024) 113–126 Y. Dou State transition probability Loop Ant population Modify tabu table Total number of ants 0.3, the path length increased rapidly. Therefore, in small-scale environments, with a PTP of around 0.3, the model could balance the needs of priority and path length. In medium scale regions, when PTP=0.5, the path growth of the model was relatively slow. In large-scale regions, the difference in path growth between any two points was relatively small. At this time, priority conditions should be considered first, and the optimal PTP value was 0.8. This study analyzed the rescue effect with and without TBS, as shown in Figure 9. Effect Analysis of Adaptive Ant Colony Algorithm with QoS… Informatica 48 (2024) 113–126 123 1 2 3 4 5 6 7 8 9 10 12.0 12.3 12.6 12.9 13.1 Number of experiments 12.2 12.1 12.4 12.5 12.7 12.8 13.0 Length of TSP /km 13.2 13.3 No temporary base station Temporary base station No temporary base station maximum Temporary base station maximum Figure 9: The impact of TBS on S/R efficiency Figure 9 shows that for drones S/R without TBS. The average path length reached 12.8km. The drones with TBS S/R had an average path of 11.4km. This indicates that the size of PTP is related to the scope of the collection area. The setting of TBS can effectively reduce the length of S/R paths and improve S/R efficiency. This study constructed a S/R map based on the data in Table 1, and the path simulation results at PTP=0.3 are displayed in Figure 10. 1000 2000 3000 4000 5000 500 1000 1500 2000 2500 3000 3500 4000 X Y (a) UAV path map 0 10 20 30 50 1.6 1.8 2.0 2.1 2.2 2.3 2.4 2.5 Number of experiments Path length (b) Path change diagram 40 1.7 1.9 Average length Shortest Length Figure 10: Trend of path changes in small-scale S/R areas Figure 10 (a) shows the motion trajectory of drones in a S/R environment. Due to the small scale of the S/R environment, sparse topology, and large distance in the Xiangling area, drones ensured priority access to areas with higher priority while also prioritizing the shortest distance factor to a greater extent. Figure 10 (b) shows the adjustment of path length during the operation of the QSACO algorithm. The average S/R length of the algorithm was 2.06km, and the shortest path length was 1.65km. The results indicate that in small-scale S/R scenarios, the QSACO algorithm has a shorter S/R path, reduces the S/R distance of drones, and improves the S/R 124 Informatica 48 (2024) 113–126 Y. Dou efficiency of drones. To further verify the progressiveness of QSACO algorithm, this study used Cuckoo Search Algorithm (CSA) for comparative analysis, as shown in Figure 11 [20]. 1 2 3 4 5 6 7 8 9 10 2.5 2.7 2.9 3.1 Length of TSP /km QSACO CAS Number of experiments (a) 32 scale 1 2 3 4 5 6 7 8 9 10 6.5 6.8 7.1 8.0 Length of TSP /km QSACO CAS Number of experiments (b) 128 scale 7.4 7.7 1 2 3 4 5 6 7 8 9 10 27 29 31 32 Length of TSP /km QSACO CAS Number of experiments (c) 512 scale 28 30 33 Figure 11: TDL under different scale scenarios Figures 11 (a), (b), and (c) show the TDL results in small, medium, and large-scale S/R scenarios. In Figure 11 (a), CAS and QSACO had similar performance effects, with a minimum TDL of 2.72km for both algorithms. In Figure 11 (b), the minimum TDL of QSACO was 6.83km, and the minimum TDL of CAS was 7.25km. In Figure 11 (c), the minimum TDL of QSACO was 28.5km, and the minimum TDL of CAS was 29.6km. The data shows that research methods can effectively find the shortest path and avoid duplicate paths. In small-scale S/R environments, research methods and advanced methods have similar performance. With the expansion of S/R scale, research methods also have certain advantages, so the proposed QSACO method has certain feasibility and reliability. 0 1500 3000 4500 6000 0 1000 2000 3000 4000 5000 6000 X Y (a) Search and rescue scenario MP-DS TI-DCN MILP QSACO Algorithm Coverage rate/% (b) Coverage result TM-LG 82 84 86 88 90 98 96 94 92 MP-DS TI-DCN MILP QSACO Algorithm Time/s (d) Search path planning time result TM-LG 0.2 0.3 0.4 0.5 0.6 1.0 0.9 0.8 0.7 MP-DS TI-DCN MILP QSACO Algorithm Path length/km (c) Search path length results TM-LG 40 41 42 43 44 48 47 46 45 P<0.05 P<0.05 P<0.05 P>0.05 P<0.05 P<0.05 P<0.05 P<0.05 P>0.05 P<0.05 P<0.05 P<0.05 Figure 12: Search and rescue results of different algorithms Effect Analysis of Adaptive Ant Colony Algorithm with QoS… Informatica 48 (2024) 113–126 125 To evaluate the research method more comprehensively, the study used a more diverse set of disaster scenarios for S/R validation. The simulated disaster scenario was shown in Figure 12(a). In Figure 12(a), there were a total of 32 scenic spots with S/R sites. S/R evaluation index adopted coverage rate, S/R path length and S/R path planning completion time. The obtained index data were analyzed using SPSS22.0 statistical software and the comparison algorithms were MP-DS, TI-DCN, MILP and TM-LG. The results were shown in Figure 12(b) to Figure 12(d). Figure 12(b) showed that the coverage rates of MP-DS, TI-DCN, MILP and TM-LG algorithms were 87.50%, 84.38%, 87.50% and 90.63%, respectively, and the coverage rate of QSACO was 96.88%. Among them, MP-DS, TI-DCN and MILP had statistically significant differences compared to QSACO algorithm (P<0.05). In Figure 12(c), from the analysis of S/R path length, the search path length of MP-DS, TI-DCN, MILP and TM-LG algorithms was 43.27km, 44.17km, 43.87km and 46.01km, respectively, and the search path length of QSACO was 40.56km. The search path length between the above four methods and QSACO algorithm had statistical significance (P<0.05). In Figure 12(d), from the analysis of the completion time of S/R path planning, the time of MP-DS, TI-DCN, MILP and TM-LG was 0.62s, 0.74s, 0.91s and 0.84s, respectively, and the time of QSACO was 0.54s. The difference of search path planning time between TI-DCN, MILP and TM-LG algorithm and QSACO algorithm was statistically significant (P<0.05). From the above simulation results, the search path planning time of QSACO algorithm was the shortest. Therefore, this paper analyzed the complexity of QSACO algorithm. In the matrix that stores the pheromone concentration on each path, its magnitude was proportional to the square of the number of nodes, denoated as (O(m2)). The path of each ant needed to be recorded, and in the worst case, the length of this record was equal to the number of nodes, i.e. (O(m)). Thus, all ant path records required (O(n\times m)) space. The amount of space required to store the QoS parameters associated with each node or path depends on the number and complexity of QoS constraints. In summary, the QSACO algorithm's spatial complexity can be expressed as approximately (O(m2 + n \times m)), and its algorithmic complexity is dependent on the rescue point and number of iterations, resulting in low complexity. 5 Discussion According to the above experimental results, the high coverage of QSACO algorithm can be attributed to its optimized search strategy and efficient path planning. By introducing QoS constraints, QSACO algorithm can not only ensure S/R quality, but also optimize S/R paths to ensure that more areas are covered. In addition, the adaptive properties of the ACO algorithm enable it to adjust search strategies based on real-time feedback, further improving the coverage of S/R missions. QSACO algorithm performs better than other algorithms in S/R path length, mainly due to its improvement of pheromone update mechanism and effective use of QoS constraints. The optimal configuration of pheromone volatility coefficient and pheromone increase intensity coefficient enables the algorithm to effectively explore and strengthen those short paths and high QoS performance paths. This not only improves the optimization speed of the path, but also avoids premature convergence to the local optimal solution, thus reducing the length of the S/R path as a whole. The superiority of QSACO algorithm in the completion time of path planning indicates its efficient computing power and fast decision-making mechanism. The algorithm achieves fast path search and decision making by precisely adjusting ant population and information heuristic factors. This increase in efficiency, especially in the face of complex S/R scenarios, significantly reduces S/R preparation time and enables UAVs to be deployed to S/R sites more quickly. In short, through the improvement of ACO algorithm and the effective integration of QoS constraints, QSACO algorithm can flexibly respond to changes in various S/R scenarios, and the introduction of QoS constraints ensures the dual guarantee of efficiency and quality of S/R tasks. Additionally, the algorithm's improvement considers the drone's ability to dynamically adjust in real-time. This allows the S/R system to promptly respond to changes in the disaster area environment, optimize S/R path planning, and ultimately accomplish efficient and precise S/R tasks. 6 Conclusion Traditional drones S/R systems have certain limitations in terms of efficiency and quality in response to complex and ever-changing disaster environments. Therefore, this study proposed a drones S/R model based on QSACO by constructing a disaster area drones S/R scenario and simulating the ACO process while introducing QoS constraints. This study validated the performance of the proposed algorithm through simulation using S/R areas of different scales. The simulation results showed that in scales of 32, 128, and 512, the QSACO achieved the shortest TDL of 1.55km, 5.98km, and 26.03km, respectively, with significant improvement compared to the ACO and PSO. In the comparative experiment with the CAS algorithm, the minimum TDL of QSACO algorithm was similar to or even better than the CAS. In summary, the proposed QSACO effectively met the path planning requirements in drones S/R scenarios in disaster areas, avoided duplicate path planning, and could solve local optimal solutions. Although the simulation results are good, there are still some limitations. Simulation experiments were based on hypothetical disaster environments, which might not fully capture the complexity and unpredictability of real disaster scenarios. For example, there may be dynamically changing obstacles, unstable weather conditions, complex terrain 126 Informatica 48 (2024) 113–126 Y. Dou and other factors in the real disaster environment, which may affect the efficiency of UAV S/R and the accuracy of path planning. When conducting S/R in disaster areas, it is important to plan a path that covers the affected area as quickly as possible. In addition to this, the drones may also need to perform various tasks such as locating and rescuing injured individuals and delivering emergency supplies. These tasks may require further adjustment and optimization of the ACO algorithm. To improve the accuracy of simulation experiments, future research should build simulation models by collecting more realistic environmental data, so as to better simulate the challenges that may be encountered in real disaster environments. Due to the diversity of S/R tasks, the algorithm design should also consider how to combine path planning with task scheduling strategies such as casualty rescue and material delivery, so as to achieve a more comprehensive S/R task planning. References [1] J. Tanwar, S. K. Sharma, M. Mittal, “Designing obstacle's map of an unknown place using autonomous drone navigation and web services,” International Journal of Pervasive Computing and Communications, vol. 19, no. 1, pp.154-169, 2023. https://doi.org/10.1108/IJPCC-07-2020-0085 [2] F. H. K. Zaman, N. M. Tahir, Y. M. Yusoff, N. M. Thamrin, and A. H. Hasmi, “Human detection from drone using you only look once (YOLOv5) for search and rescue operation,” Journal of Advanced Research in Applied Sciences and Engineering Technology, vol. 30, no. 3, pp. 222-235, 2023. https://doi.org/10.37934/araset.30.3.222235 [3]J. Horyna, T. Baca, V. Walter, D. Albani, D. Hert, E. Ferrante, and M. Saska, “Decentralized swarms of unmanned aerial vehicles for search and rescue operations without explicit communication,” Autonomous Robots, vol. 47, no. 1, pp. 77-93, 2023. https://doi.org/10.1007/s10514-022-10066-5 [4]A. Depold,C. Dorn, S. Erhardt, R. Weigel, and F. Lurz, “A direction-of-arrival estimation system for UAV-assisted search and rescue: locating mobile phones to improve the survival chance of disaster victims,” IEEE Microwave Magazine, vol. 24, no. 3, pp. 59-64, 2023. https://doi.org/10.1109/MMM.2022.3226548 [5]A. Albanese, V. Sciancalepore, and X. Costa-P érez, “SARDO: An automated search-and-rescue drone-based solution for victims localization,” IEEE Transactions on Mobile Computing, vol. 21, no. 9, pp. 3312-3325, 2021. https://doi.org/10.1109/TMC.2021.3051273 [6]J. Dong, K. Ota, and M. Dong, “UAV-based real-time survivor detection system in post-disaster search and rescue operations,” IEEE Journal on Miniaturization for Air and Space Systems, vol. 2, no. 4, pp. 209-219, 2021. https://doi.org/10.1109/JMASS.2021.3083659 [7]R. G. Ribeiro, L. P. Cota, T. A. M. Euz ébio, J. A. Ram í rez, and F. G. Guimarães, “Unmanned-aerial-vehicle routing problem with mobile charging stations for assisting search and rescue missions in postdisaster scenarios,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 52, no. 11, pp. 6682-6696, 2021. https://doi.org/10.1109/TSMC.2021.3088776 [8]M. Atif, R. Ahmad, W. Ahmad, L. Zhao, and J. Rodrigues, “UAV-assisted wireless localization for search and rescue,” IEEE Systems Journal, vol. 15, no. 3, pp. 3261-3272, 2021. https://doi.org/10.1109/JSYST.2020.3041573 [9]G. Li, and Y. Li, “UAV path planning based on improved ant colony algorithm,” Second International Conference on Algorithms, Microchips, and Network Applications (AMNA 2023). SPIE, vol. 12635, pp. 59-63., 2023. https://doi.org/10.1117/12.2678893 [10] D. Mandloi, R. Arya, and A. K. Verma, “Unmanned aerial vehicle path planning based on A* algorithm and its variants in 3d environment,” International Journal of System Assurance Engineering and Management, vol. 12, no. 1, pp. 990-1000, 2021. https://doi.org/10.1007/s13198-021-01186-9 [11]F. Gul, W. Rahiman, S. S. N. Alhady, A. Ali, I. Mir, and A. Jalil, “Meta-heuristic approach for solving multi-objective path planning for autonomous guided robot using PSO – GWO optimization algorithm with evolutionary programming,” Journal of Ambient Intelligence and Humanized Computing, vol. 12, no. 1, pp. 7873-7890, 2021. https://doi.org/10.1007/s12652-020-02514-w [12]W. Chi, Z. Ding, J. Wang, G. Chen, and L. Sun, “A generalized Voronoi diagram-based efficient heuristic path planning method for RRTs in mobile robots,” IEEE Transactions on Industrial Electronics, vol. 69, no. 5, 4926-4937, 2021. https://doi.org/10.1109/TIE.2021.3078390 [13]Y. Tan, J. Ouyang, Z. Zhang, Y. Lao, and P. Wen, “Path planning for spot welding robots based on improved ant colony algorithm,” Robotica, vol. 41, no. 3, 926-938, 2023. https://doi.org/10.1017/S026357472200114X [14]S. Choudhuri, S. Adeniye, and A. Sen, “Distribution alignment using complement entropy objective and adaptive consensus-based label refinement for partial domain adaptation,” Artificial Intelligence and Applications. vol. 1, no. 1, pp. 43-51, 2023. https://doi.org/10.47852/bonviewAIA2202524 [15]X. Meng, X. Zhu, and J. Zhao, “Obstacle avoidance path planning using the elite ant colony algorithm for parameter optimization of unmanned aerial vehicles,” Arabian Journal for Science and Engineering, vol. 48, no. 2, pp. 2261-2275, 2023. https://doi.org/10.1007/s13369-022-07204-7 [16]X. Wang, and X. Song, “Optimal path planning of Effect Analysis of Adaptive Ant Colony Algorithm with QoS… Informatica 48 (2024) 113–126 127 logistics distribution of urban and rural agricultural products from the perspective of supply chain,” Informatica, vol. 47, no. 5, pp. 69-74, 2023. https://doi.org/10.31449/inf.v47i5.4557 [17]D. Chen, X. M. You, and S. Liu, “Ant colony algorithm with n- α -measure and migration learning,” Arabian Journal for Science and Engineering, vol. 48, no. 2, pp. 1873-1890, 2023. https://doi.org/10.1010.1007/s13369-022-07076-x [18]X. Wang, Z. Guo, H. Zhang, C. Wang, and Y. Wang, “Snowmelt detection on the Antarctic ice sheet surface based on XPGR with improved ant colony algorithm,” International Journal of Remote Sensing, vol. 44, no. 1, pp. 142-156, 2023. https://doi.org/10.1080/01431161.2022.2161851 [19]Y. Gao, H. Wu, and W. Wang, “A hybrid ant colony optimization with fireworks algorithm to solve capacitated vehicle routing problem,” Applied Intelligence, vol. 53, no. 6, pp. 7326-7342, 2023. https://doi.org/10.1007/s10489-022-03912-7 [20]M. Morin, I. Abi-Zeid, and C. G. Quimper, “Ant colony optimization for path planning in search and rescue operations,” European Journal of Operational Research, vol. 305, no. 1, pp. 53-63, 2023. https://doi.org/10.1016/j.ejor.2022.06.019