SÁNG THỨ BẢY, 26/11/2022
Phòng: F202
Phiên 1: Lý thuyết tối ưu
Chủ tọa: PGS.TS. Nguyễn Lê Hoàng Anh

Phiên 2: Lý thuyết tối ưu
Chủ toạ TS. Võ Sĩ Trọng Long

CHIỀU THỨ BẢY, 26/11/2022
Phòng: F.202
Phiên 3: Cơ học
Chủ toạ: TS. Vũ Đỗ Huy Cường 

Phiên 4: Tin học ứng dụng

Chủ toạ: PGS.TS. Nguyễn Thanh Bình

Dinh Vinh Hien
SCALARIZATION IN SET OPTIMIZATION WITH VARIABLE ORDERING STRUCTURES AND APPLICATIONS

In this talk, we consider set optimization problems with variable ordering structures. We propose the new versions of scalarization functions for the reference problems. Under the appropriate assumptions, we study the properties of such functions. As applications of the above results, the optimality conditions for solutions of the set optimization problems are established. Our results generalize or extend the existing ones in the literature.

Mai Van Duy and Nguyen Minh Tung
KARUSH-KUHN-TUCKER CONDITIONS AND DUALITY FOR A CLASS CONVEX ADJUSTABLE ROBUST OPTIMIZATION PROBLEM

Robust optimization (RO) is a popular methodology to solve mathematical optimization problems with uncertain data. For the RO problem, the solutions that are immune to all perturbations of the data in a so-called uncertainty set are found. Adjustable robust optimization (ARO), on the other hand, is a branch of RO where some of the decision variables can be adjusted after the uncertain data (or some of its portion) reveals itself. This work deal with the optimality conditions and duality for a class of convex adjustable robust optimization problems.

Bảo Nguyễn Xuân Duy and Tùng Nguyễn Minh
NEW SECOND-ORDER LIMITING DIRECTIONAL DERIVATIVES AND OPTIMIZATION

In this paper, new second-order set-valued directional derivatives are proposed for C$^1$ functions, whose derivative is locally calm (stable), in normed spaces. Its existence, main calculus, as well as Taylor's expansions are studied. We then employ them to investigate optimality conditions for optimization problems with geometric and functional constraints. The results also improve the corresponding ones for problems involving C$^{1,1}$ functions. Examples that analyze and illustrate our results are given.

Tung Le Thanh and Khai Tran Thien
LAGRANGE DUALITY AND SADDLE POINT OPTIMALITY CONDITIONS
FOR MULTIOBJECTIVE SEMI-INFINITE PROGRAMMING WITH VANISHING CONSTRAINTS

The objective of this talk is to investigate multiobjective semi-infinite programming problems with vanishing constraints (MSIPVC). We firstly formulate the scalarized Lagrange dual problems for MSIPVC and explore duality relations under convexity assumptions. Then, the saddle point optimality conditions for MSIPVC are considered. Some examples are also given to illuminate the main results.

Ba Khiet Le
SLIDING MODE OBSERVERS FOR TIME-DEPENDENT SET-VALUED LUR'E SYSTEMS SUBJECT TO UNCERTAINTIES

Designing observers for dynamical systems plays an important role in\nthe modern control theory due to the lack of full information in measured outputs.\nThe current paper proposes a sliding mode observer for a general class of\nLur'e systems subject to uncertainties where feedbacks involve time-dependent\nset-valued mappings. To the best of our knowledge, sliding mode observers for setvalued\nLur'e systems, even for the
simple static case, have not been considered in\nthe literature. Exponential convergence of the observer state and nite-time convergence\nof the output estimation error are guaranteed without using any linear\ntransformations. In addition, our design can also deduce H^∞ observers.

Phạm Thanh Dược
BIỂU DIỄN VÔ HƯỚNG VÀ TÍNH LIÊN TỤC HAUSDORFF CỦA ÁNH XẠ NGHIỆM
CHO BÀI TOÁN TỐI ƯU HÓA TẬP CHỨA THAM SỐ

Trong báo cáo này, chúng tôi đề xuất các hàm vô hướng phi tuyến mới và sử dụng chúng để hình thành các biểu diễn và điều kiện ổn định cho các bài toán tối ưu hóa tập chứa tham số. Chính xác hơn, trước tiên chúng tôi giới thiệu các hàm vô hướng phi tuyến mới cho các tập hợp và thảo luận về các thuộc tính của chúng. Sau đó, chúng tôi cung cấp các biểu diễn vô hướng của tập nghiệm bài toán tối ưu tập. Cuối cùng, chúng tôi nghiên cứu các điều kiện đủ cho sự liên tục Hausdorff của ánh xạ nghiệm gần đúng với các bài toán tối ưu tập chứa tham số không đơn điệu và không lồi.

Thành Tài Võ
ON THE STABILITY FOR NONLINEAR MULTIOBJECTIVE OPTIMAL CONTROL PROBLEMS

In this talk, we consider nonlinear multiobjective optimal control problems and study the stability of efficient solutions for such problems. We first introduce a generalized boundedness condition and discuss its relations with some typical existing ones in the literature. Next, we investigate upper and lower semicontinuity properties of the constraint maps. Finally, under suitable assumptions, sufficient conditions for the convergence of efficient solutions are established.

Huu Danh Nguyen and Ngoc Tam Tran
QUALITATIVE PROPERTIES OF SOLUTIONS TO SET OPTIMIZATION PROBLEMS

In this report we consider set optimization problems involving set relations. Firstly, utilizing the well-known KKM-Fan Lemma and relaxed convexity assumptions, we study existence conditions for these problems. Secondly, we introduce parametric nonlinear scalarization functions for sets and study their properties. By using the concerning functions, we investigate relationships between set optimization problems and equilibrium problems. Finally, sufficient conditions for the H\"older continuity of solution maps to such problems via equilibrium problems are established.

Huynh Thanh Toan, Avner Peleg and Nguyen Minh Quan
IMPACT OF TWO-PHOTON ABSORPTION ON FAST TWO-BEAM COLLISIONS IN LINEAR OPTICAL MEDIA

In this talk, we present the impact of two-photon absorption on fast two-beam collisions in linear bulk optical media in spatial dimension higher than 1. We show that the collision leads to a change in the beam shapes in the direction transverse to the relative velocity vector. The theoretical predictions are confirmed by extensive numerical simulations with the linear propagation model with weak cubic loss.

Vu Do Huy Cuong and Nguyen Ngoc Minh
TOPOLOGY OPTIMIZATION OF STRUCTURE USING PROPORTIONAL TOPOLOGY OPTIMIZATION METHOD.

Proportional Topology Optimization (PTO) is a non-gradient topology optimization method which is simple to understand, easy to implement, and is also efficient and accurate at the same time. This method has just appeared but has achieved certain achievements in comparison with the others. In this work, the PTO minimum compliance algorithm is used to find the optimal shape of structure’s body under effect of heat or elastic mechanics. The solutions is very interesting. Some numerical examples are showed and compared to given results.

Vu Do Huy Cuong, Nguyen Thoi Trung and Ly Khuong Duy
THE MULTI-PHYSICAL ZIG-ZAG MITC4 ELEMENT FOR ANALYZING FREE VIBRATION
AND STATIC BEHAVIOR OF PIEZOELECTRIC/PIEZOMAGNETIC COMPOSITE THIN-WALLED STRUCTURES

This study proposed an efficient numerical approach for static response and free vibration analysis of piezoelectric/piezomagnetic composite thin-walled intelligent structures. The coupled multi-physical MITC4 element based on zig-zag theory is used to assess the free vibration and static response of multilayered/functionally graded magneto-electro-elastic structures. The coupled response of the MEE structure is also considered due to the coupling characteristic between magnetic, electric, and elastic fields. Several examples are given to show the high reliability of the proposed formulation.

Pham Van Nha, Ngo Thanh Long and Pham The Long
INTERVAL-VALUED FUZZY THRESHOLDING APPROACH FOR IMAGE SEGMENTATION

Fuzzy sets were researched to solve complex problems where there are many ambiguous and uncertain factors. Type-2 fuzzy sets have been proved to be remarkably effective in some scientific fields such as pattern recognition, machine learning, image processing, etc. However, Type-2 fuzzy sets are limited in computing complexity. Interval-valued fuzzy sets are the special case of Type-2 fuzzy sets to overcome the limitations of traditional fuzzy sets. In this paper, we have proposed a new thresholding algorithm by interval-valued fuzzy sets approach. Firstly, interval-valued fuzzy sets are used to construct new objective function of the weak continuity constraints between thresholds.
Secondly, the global energy function is calculated to determine the optimal threshold in the considerated threshold range. Thirdly, the local energy function of the segments is calculated to determine which segment will continue to be thresholded. Finally, the stop condition of the proposed algorithm is calculated to end the thresholding process and produce the segment results. Experiments have been conducted on some gray multi-level images. Experimental results have demonstrated the outstanding performance of the proposed algorithm compared to some previous proposed techniques.

Do Tran Ngoc and Hung Nguyen Van
EFFICIENCY MODEL ON INSECT RECOGNITION USING CONVOLUTIONAL NEURAL NETWORK

Foods play a bigger part in human existence than ever before, and planning food crops is just as crucial as rising food demand. The major cause of the productivity decline in food crops is insect pests. Accurate insect identification is more necessary because there is just one approach available for eliminating a single or a group of problem insects. Insect pest species are numerous, as are the individuals within each species. Because it requires a lot of labor, the problem of pest identification and categorization is difficult for farmers and scientists who try to discover the best answer. When computers get quicker and computer vision reaches its pinnacle, manual labour must be replaced by an automated system for pest recognition and statistics. Our solution uses deep learning with a convolutional neural network and focuses on the pre-processing stage, which is based on the open- source large-scale dataset IP102. The evaluation of our model yielded positive results, with the classification issue accuracy exceeding 60\\% and the mAP detection rate
reaching 50\\%.

Tram Vo Thi Bich, Tram Vo Thi Bich and Tram Vo Thi Bich
XÁC ĐỊNH THỜI GIAN CHUYỂN GIAO TỪ PHƯƠNG TIỆN ĐẾN CƠ SỞ HẠ TẦNG TRONG GIAO TIẾP BẰNG ÁNH SÁNG NHÌN THẤY

Trong hệ thống giao tiếp ánh sáng nhìn thấy giữa phương tiện xe với cơ sở hạ tầng (vehicle-to-infrastructure visible light communication- V2IVLC), phương tiện nhận dữ liệu từ đèn đường LED được đặt dọc theo hai bên đường. Tại một thời điểm nhất định, một chiếc xe chỉ giao tiếp với một nhóm đèn LED. Khi xe di chuyển, xe cần chuyển giao tiếp từ nhóm đèn LED hiện tại sang nhóm tiếp theo. Do phương tiện di chuyển nhanh và độ phủ của từng nhóm LED nhỏ nên việc bàn giao giữa các nhóm LED là một bài toán khó. Bài báo này đề xuất một quy trình chuyển giao tiếp giữa các nhóm đèn LED cho một hệ thống V2IVLC.

.

.