• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.7
• /
• pp.263-270
• /
• 2021

Reinforcement learning(RL) is the method to find an optimal policy through training. and it is one of popular methods for solving lifesaving and disaster response problems effectively. However, the conventional reinforcement learning method for disaster response utilizes either simple environment such as. grid and graph or a self-developed environment that are hard to verify the practical effectiveness. In this paper, we propose the design of a disaster response RL environment which utilizes the detailed property information of the disaster simulation in order to utilize the reinforcement learning method in the real world. For the RL environment, we design and build the reinforcement learning communication as well as the interface between the RL agent and the disaster simulation. Also, we apply the dimension reduction method for converting non-image feature vectors into image format which is effectively utilized with convolution layer to utilize the high-dimensional and detailed property of the disaster simulation. To verify the effectiveness of our proposed method, we conducted empirical evaluations and it shows that our proposed method outperformed conventional methods in the building fire damage.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.8 no.3
• /
• pp.1-8
• /
• 1998

This paper proposes a reinforcement genetic programming based on the reinforcement learning method for the performance improvement of genetic programming. Genetic programming which has tree structure program has much flexibility of problem expression because it has no limitation in the size of chromosome compared to the other evolutionary algorithms. But worse results on the point of convergence associated with mutation and crossover operations are often due to this characteristic. Therefore the sizes of population and maximum generation are typically larger than those of the other evolutionary algorithms. This paper proposes a new method that executes crossover and mutation operations based on reinforcement and inhibition mechanism of reinforcement learning. The validity of the proposed method is evaluated by appling it to the artificial ant problem.

• Structural Engineering and Mechanics
• /
• v.5 no.5
• /
• pp.565-576
• /
• 1997

The material-based homogenization design method generates arbitrary topologies of initial structural design as well as reinforcement structural design by controlling the amount of material available. However, if a small volume constraint is specified in the design of Lightweight structures, thin and slender structures are usually obtained. For these structures stability becomes one of the most important requirements. Thus, to prevent overall buckling (that is, to increase stability), the objective of the design is to maximize the buckling load of a structure. In this paper, the buckling analysis is restricted to the linear buckling behavior of a structure. The global stability requirement is defined as a stiffness constraint, and determined by solving the eigenvalue problem. The optimality conditions to update the design variables are derived based on the sequential convex approximation method and the dual method. Illustrated examples are presented to validate the feasibility of this method in the design of structures.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2005.11a
• /
• pp.17-20
• /
• 2005

Moment-curvature relation of RC pier is influenced greatly in occurrence form of crack and difference is happened according to consideration existence and nonexistence of tension stiffening effect. However, studies considering these is very insufficient misgovernment. Also, it is sometimes unavoidable lap splice of axial reinforcement in plastic hinge region of RC piers. However, specific design standard about lap splice of axial reinforcement is unprepared real condition and study about effect that lap splice of axial reinforcement get in occurrence form of crack is insufficient misgovernment. Therefore, in this paper, experiments are performed with hollow RC piers that do lap splice of axial reinforcement by main variable. And this study present analytical method about moment-curvature relation of hollow RC pier that consider tension stiffening effect and analyze effect that lap splice of axial reinforcement gets in occurrence form of crack. Analytic method of moment-curvature relation of RC pier that present in this study shows very similar motion with experiment result and crack interval of RC pier is suffering dominate impact in the augmented reinforcement amount by lap splice and average crack interval decreases as lap splice ratio increases.

• Computers and Concrete
• /
• v.27 no.4
• /
• pp.385-393
• /
• 2021

Reinforcement corrosion is the main cause of the durability failure of reinforced concrete (RC) structure. In this paper, a three-dimensional (3D) numerical model of macro-cell corrosion is established to reveal the corrosion mechanisms of steel reinforcement in RC structure. Modified Direct Iteration Method (MDIM) is employed to solve the system of partial differential equations for reinforcement corrosion. Through the sensitivity analysis of electrochemical parameters, it is found that the average corrosion current density is more sensitive to the change of cathodic Tafel slope and anodic equilibrium potential, compared with the other electrochemical parameters. Furthermore, both the anode-to-cathode (A/C) ratio and the anodic length have significant influences on the average corrosion current density, especially when A/C ratio is less than 0.5 and anodic length is less than 35 mm. More importantly, it is demonstrated that the corrosion rate of semi-circumferential corrosion is much larger than that of circumferential corrosion for the same A/C ratio value. The simulation results can give a unique insight into understanding the detailed electrochemical corrosion processes of steel reinforcement in RC structure for application in service life prediction of RC structures in actual civil engineer.

• Journal of the Korea Concrete Institute
• /
• v.22 no.4
• /
• pp.527-534
• /
• 2010

This study presents a simple method to improve the bond performance of reinforced concrete (RC) beams having high-strength shear reinforcement. In general, the yield strength and the ratio of shear reinforcements are the main parameters governing the shear capacity of RC beams. The yield strength of shear reinforcement, however, has little influence on the bond capacity of RC beams. Therefore, a sudden bond failure of the members with high-strength shear reinforcement can occur before flexural failure. To estimate the structural performance of the proposed method, four RC beams were cast and tested. The main test parameters were the yield strength, ratio, and reinforcing types of shear reinforcements. The experimental results indicated that the proposed method was able to effectively improve the bond performance of RC beams with high-strength shear reinforcement.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.1
• /
• pp.1-5
• /
• 2005

The purpose of reinforcement learning is to maximize rewards from environment, and reinforcement learning agents learn by interacting with external environment through trial and error. Q-Learning, a representative reinforcement learning algorithm, is a type of TD-learning that exploits difference in suitability according to the change of time in learning. The method obtains the optimal policy through repeated experience of evaluation of all state-action pairs in the state space. This study chose n-Queen problem as an example, to which we apply reinforcement learning, and used Q-Learning as a problem solving algorithm. This study compared the proposed method using reinforcement learning with existing methods for solving n-Queen problem and found that the proposed method improves the convergence rate to the optimal solution by reducing the number of state transitions to reach the goal.

• Journal of Internet Computing and Services
• /
• v.9 no.6
• /
• pp.165-177
• /
• 2008

An important issue in Multiagent reinforcement learning is how an agent should learn its optimal policy in a dynamic environment where there exist other agents able to influence its own performance. Most previous works for Multiagent reinforcement learning tend to apply single-agent reinforcement learning techniques without any extensions or require some unrealistic assumptions even though they use explicit models of other agents. In this paper, a Naive Bayesian based policy model of the opponent agent is introduced and then the Multiagent reinforcement learning method using this model is explained. Unlike previous works, the proposed Multiagent reinforcement learning method utilizes the Naive Bayesian based policy model, not the Q function model of the opponent agent. Moreover, this learning method can improve learning efficiency by using a simpler one than other richer but time-consuming policy models such as Finite State Machines(FSM) and Markov chains. In this paper, the Cat and Mouse game is introduced as an adversarial Multiagent environment. And then effectiveness of the proposed Naive Bayesian based policy model is analyzed through experiments using this game as test-bed.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2010.09a
• /
• pp.303-314
• /
• 2010

This study, as basic research which was intended to develope the surface reinforcement method using reinforcement material which is applicable to very soft ground in Korea, was aimed at proposing the design parameter for the surface ground improvement method. To that end, a wide width tensile test using geotextile, geogrid and steel bar (substitute for bamboo) and 49 kinds of the laboratory model tests were conducted. And the result the study suggested $\beta_s$, the stiffness coefficient to evaluate the stiffness effect of reinforcement materials. Then, it was also found that the stiffness coefficient, $\beta_s$ as the testing constant would be appropriate as high as 1.0, 1.1 and 1.5 for geotextile, geogrid and steel bar, respectively. And It was evaluated that the stiffness effect affecting reinforcement improvement effect would be reduced as the thickness of embeded depth increases and that RFe, the stiffness effect reduction coefficient would have positive correlation with H/B. Finally, it was confirmed that the bearing capacity gained from the method to calculate bearing capacity, which was suggested in the study, would almost correctly estimate the capacity, demonstrating the appropriateness of the proposed bearing capacity calculation method.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2009.03a
• /
• pp.677-685
• /
• 2009

From the result of precise field investigation and stability examination for the cut slope, following results were acquired. 1. The cause of the cut slope collapse seems that it came from complex actions such as the weathering rock itself, existing fault zone and crack from excessive blasting. 2. As a result of analysis, it is appeared that the cut slope can be in danger of plane destruction. 3. The reinforcement force is decided by the result of limit equilibrium analysis. 4. For reinforcement the method after relaxing the slope was judged as the most proper method to the cut slope as comparing/analyzing Rock Anchor, Rock Bolt and method after relaxing the slope.