Search the digital library catalog

Abstract: Traveling Salesman Problems (TSPs) have been a long-lasting interesting challenge to researchers in different areas. The difficulty of such problems scales up further when multiple objectives are considered concurrently. Plenty of work in evolutionary algorithms has been introduced to solve multi-objective TSPs with promising results, and the work in deep learning and reinforcement learning has been surging. This paper introduces a multi-objective deep graph pointer network-based reinforcement learning (MODGRL) algorithm for multi-objective TSPs. The MODGRL improves an earlier multi-objective deep reinforcement learning algorithm, called DRL-MOA, by utilizing a graph pointer network to learn the graphical structures of TSPs. Such improvements allow MODGRL to be trained on a small-scale TSP, but can find optimal solutions for large scale TSPs. NSGA-II, MOEA/D and SPEA2 are selected to compare with MODGRL and DRL-MOA. Hypervolume, spread and coverage over Pareto front (CPF) quality indicators were selected to assess the algorithms’ performance. In terms of the hypervolume indicator that represents the convergence and diversity of Pareto-frontiers, MODGRL outperformed all the competitors on the three well-known benchmark problems. Such findings proved that MODGRL, with the improved graph pointer network, indeed performed better, measured by the hypervolume indicator, than DRL-MOA and the three other evolutionary algorithms. MODGRL and DRL-MOA were comparable in the leading group, measured by the spread indicator. Although MODGRL performed better than DRL-MOA, both of them were just average regarding the evenness and diversity measured by the CPF indicator. Such findings remind that different performance indicators measure Pareto-frontiers from different perspectives. Choosing a well-accepted and suitable performance indicator to one’s experimental design is very critical, and may affect the conclusions. Three evolutionary algorithms were also experimented on with extra iterations, to validate whether extra iterations affected the performance. The results show that NSGA-II and SPEA2 were greatly improved measured by the Spread and CPF indicators. Such findings raise fairness concerns on algorithm comparisons using different fixed stopping criteria for different algorithms, which appeared in the DRL-MOA work and many others. Through these lessons, we concluded that MODGRL indeed performed better than DRL-MOA in terms of hypervolumne, and we also urge researchers on fair experimental designs and comparisons, in order to derive scientifically sound conclusions.
Keywords: multi-objective optimization, traveling salesman problems, deep reinforcement learning
Published in DKUM: 28.03.2024; Views: 184; Downloads: 32
Full text (7,89 MB)
This document has many files! More...

Abstract: This article details how the serviceability of a flexural pavement structure is improved by incorporating the Cellular Confinement System (geocells). For this purpose, two different pavement structures, with and without embedded geocells, were manufactured in a laboratory and an accelerated traffic type of loading was applied. The vertical and horizontal cyclic loads were applied simultaneously to simulate the effect of principal stress rotation. A comparative study between the conventional flexible pavement and the geocell-reinforced flexible pavement is presented. Additionally, numerical models of the laboratory tests were built and the results were compared. The simulation of the experimental tests using the Traffic Load Simulator (TLS) are carried out using the FEM and advanced models that describe the permanent strain behavior of the unbound granular material. The development of permanent deformation within the pavement structure, with and without the geocells, is also presented through the numerical model, which was verified by the experimental results.
Keywords: flexible pavement, reinforcement, geocell, traffic load simulator, permanent deformation, numerical modeling
Published in DKUM: 18.06.2018; Views: 1313; Downloads: 104
Full text (1,39 MB)
This document has many files! More...

Abstract: The investigation on designing supporting measures for geotechnical structures consists of the proper selection of the input data for the stress – strain analysis of the excavation process and the selection of retaining measures. In the preliminary phase of investigation the area must be geologically mapped in detail, and discontinuities precisely measured, boreholes and their “in situ” tests listed and samples for laboratory examinations taken. The new method allows rock mass classification for different rock quality, from soft rocks to mixed rock masses, as well as determination of the geological strength index GSI. The limits for using the suggested method are persistent discontinuities in rock mass which lead to translation or rotation failure mechanisms, either in a single plane or as a wedge. In all cases where discontinuities play a significant role, the rock mass structure must be considered and kinematical analyses performed. In other circumstances, a rock may be uniform and reasonably isotropic due to the geometry of discontinuities and their mutual intersections. The value range of GSI is first determined in the beginning of investigation and later in the excavation phase by considering the disturbance factor D, which expresses disturbances caused by excavation methods and rock mass relaxation. The strength and deformability parameters of rocks of different quality are determined by the generalized Hoek-Brown failure criterion and applied to shallow and deep tunnels or slopes. Before the start of excavations work and after establishing the retaining measures, the analysis results are checked by monitoring. New methods include the determination of post peak strength parameters of rock mass after relaxation, and routine measurements. The newest measurement system however allows direct readings of displacements of the rock mass in both the elastic and post-peak states. With back analyses we determine the softening behavior of the rock mass and a possible need for increasing retaining measures. Such a design method enables the optimization of retaining measures and the reduction of investment costs.
Keywords: geological strength index – GSI, simulation of a triaxial large scale »in situ« test, numerical modeling, retaining measures, rock reinforcement, monitoring, back analyses, strain softening, rock mass relaxation
Published in DKUM: 16.05.2018; Views: 1325; Downloads: 69
Full text (1,05 MB)
This document has many files! More...

Abstract: This paper provides an experimental analysis of timber-framed walls, coated with carbon fibre-reinforced polymers (CFRP) strengthened fibre-plaster boards, usually used as main bearing capacity elements in the construction of prefabricated timber structures. The tensile strength of the fibre-plaster boards is lower than the strength of timber frame, therefore it is convenient to strengthen boards with high-strength materials in order to gain a higher capacity. It has been shown that the inclusion of CFRP diagonal strip reinforcement on the load-carrying capacity can be quite high and that it is maximized when the carbon strips are connected to the timber frame. On the other hand, the ductility itself was not significantly improved. The test samples proved an important distinction in behaviour in timber frame-fibreboard connecting area, dependant on the boundary conditions between inserted CFRP strips and timber frame. It has been shown that proposed simplified Eurocode 5 methods, applicable for wood-based sheathing boards, could be unsuitable for the problems presented.
Keywords: civil engineering, timber structures, panel walls, fibre-plaster boards, steel reinforcement, carbon fibres
Published in DKUM: 01.06.2012; Views: 2542; Downloads: 101
Link to full text

Abstract: The paper provides semi-analytical modelling for prefabricated timber-framed walls using the modified ▫$/gamma$▫-method. The walls are composed of a timber frame and gypsum plasterboards (GPB) which are flexibly connected to the timber frame. Because the tensile strength of the GPB is approximately 10-times lower than the compressive one, it is convenient to strengthen the boards in their tensile diagonal direction with carbon fibre-reinforced polymer (CFRP) strips, which are glued to the boards. Additionally, in the proposed mathematical model the classical beam theory is used taking into account a fasteners' flexibility in the timber frame - GPB connecting area, as well as crack appearing in the GPB and an influence of the inserted CFRP strips. Therefore the CFRP strip contribution is considered with a modified slip modulus which results in the fictive increased stiffness coefficient of the fasteners.
Keywords: civil engineering, timber structures, frame walls, fibre-plaster boards, steel reinforcement, CFRP strips, carbon fibre-reinforced polymer, mathematical modelling
Published in DKUM: 01.06.2012; Views: 2349; Downloads: 32
Link to full text