Hard Optimization Research Articles

Optimization of Hardness as a Textural Property of a Fruit Enriched Honey Sweetened Snack Bar for Children Aged 5 to 13 Years

Optimization for hardness of the snack bar formulated from selected fruits and grain amaranths bound with honey was conducted using Response Surface Methodology (RSM). To achieve a desired degree of hardness, honey content and compression force were varied. The effects of this variation was investigated to determine the best Honey: Compression force combination that results into the desired hardness. Thirteen combinations were considered. The force-honey combination varied between 9.82 to 41.87 kN and 30 to 80 g respectively. Results revealed that both honey and compression force had a significant effect (p < 0.05) on hardness of the snack bar. The optimal hardness of the snack bar was found at honey: compression force of 41.18 g (22.7%), 41.87 kN. At this optimal point, hardness was found to be 48.2 N/mm which is acceptable for children aged 5 to 13 years.

Energy efficient clustering and sink mobility protocol using Improved Dingo and Boosted Beluga Whale Optimization Algorithm for extending network lifetime in WSNs

In Wireless Sensor Networks (WSNs), the potential design challenge of energy efficiency is determined to be handled through the strategies of clustering and routing. The approaches of clustering and routing in WSNs pertain to the problems of Non-deterministic Polynomial (NP)-hard optimization. In this context, swarm intelligence-based algorithms are identified to be suitable and ideal for determining near-optimal and optimal solutions in the search space. On the other hand, APTEEN routing protocol possesses the issues that are related to unnecessary energy drain, ineffective overall network coverage and premature death of certain nodes. To address these issues, an attempt to optimize the APTEEN routing protocol using Dingo Optimization Algorithm (DOA) and Beluga Whale Optimization Algorithm (BWOA) is made in this proposed clustering protocol. With this motivation, Improved Dingo and Boosted Beluga Whale Optimization Algorithm (IDBBWOA) is proposed for determining the optimal cluster head and perform energy-efficient routing to minimized the energy consumption and maximize the lifetime of the network. It specifically used Improved Dingo Optimization Algorithm (IDOA) for attaining cluster head (CH) selection and energy efficient routing through the adoption fitness parameters of Residual Energy, Distance within and between Clusters, Network coverage, Node Degree for maximizing the rate of reliable data dissemination. It also incorporated Boosted Beluga Whale Optimization Algorithm (BBWOA) for determining the optimal points over the sink node can be moved to prevent multi-hop between CHs and the sink nodes, since it is essential for addressing the issue of hot-spot and extends the network lifetime. The simulation results of the proposed IDBBWOA approach revealed its efficacy in improving the mean throughput by 18.92 %, sustaining alive nodes by 34.28 %, and maintaining residual energy by 29.34 %, compared to the benchmarked approaches used for evaluation.

Monitoring of laser-induced fast recrystallization in SS-316L through synchrotron X-ray diffraction

Laser based heat treatments have multiple applications for local tuning of microstructures. They are particularly relevant in the context of laser-based additive manufacturing (LAM), in order to relieve residual stresses or introduce equiaxed structures through recrystallization. In 316L, high stored energy inherited from the high cooling rates of laser processing creates strong crystallographic textures and stable cellular walls, which challenges the possibility of inducing fast recrystallization. In a first step, we perform here a laser heat treatment on a highly deformed 316L steel, and show that in-situ X-ray Diffraction (XRD) is a viable tool for the monitoring of fast recrystallization. The evolution of Bragg peak width over time is connected to the evolution of dislocation density and grain size, thanks to Electron Back-Scattered Diffraction (EBSD) and Transmission Electron Microscopy (TEM) characterization. X-ray signals are shown to reveal different stages of recrystallization, in particular the development of dislocation cells and the onset of nucleation. This study also demonstrates how local heat sources such as lasers may help designing composite-like materials for grain size, texture and hardness optimization.

Filter Clustering for Compressing CNN Model With Better Feature Diversity

As a practical approach for compressing convolutional neural networks (CNNs), network pruning has been rapidly developed in recent years. The conventional methods prune inactive filters permanently from models to reduce the width of each layer and then train the pruned model until convergence. However, such methods have limitations in that: (1) The activation-based pruning criteria ignore the correlation between filters, leading to attenuation in types of features; (2) The permanent filter removal restricts the architecture of models in the subsequent training so that reducing the chances of learning more features; (3) The single-width compression may generate narrow layers that block the information flow, resulting in limited feature capacity in the next layers and hard optimization. These limitations reduce the feature diversity in the pruned model and thus lead to sub-optimal model quality. In this paper, a compression method named filter clustering is proposed to rectify the problem of poor feature diversity in traditional pruning and achieve better model quality from three perspectives. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Firstly</i> , to maintain the variety of features after pruning, we treat the model compression as a clustering task and merge filters with similar outputs, rather than removing inactive filters. Specifically, a handy estimation approach is designed to convert the similarity of the output into filter similarity, which liberates the measurement from sampling numerous images. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Secondly</i> , to increase the probability of learning more features during training, we propose a periodic training and clustering pipeline, which creates a larger optimization space by dynamically exploring different sub-model architectures. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Finally</i> , to prevent the feature capacity from being influenced by the narrow layers, we introduce and leverage a fusible anti-blocking branch to smoothly remove such layers. Extensive experiments demonstrate that the proposed method can achieve compact models with better feature diversity and reduce 1% ~ 15% more calculations than the previous methods while maintaining performance.

Graph Matching with Distance-Preserving Crossover

Graph models are fundamental to any kind of application on structured real-world problems. Any comparison between graphs by a graph distance measure requires the solution of the inexact graph matching problem, which constitutes a hard combinatorial optimization problem. An inexact matching problem includes in its formulation robustness to any type of perturbation, such as, for instance, noise, inherently present in real-world environments. In this paper, we introduce the concept of distance-preserving crossover operators for genetic algorithms for this task. For large graphs, our algorithm outperforms any state-of-the-art approximate algorithm—in particular, genetic algorithms with alternative crossover operators, which are to the best of our knowledge currently limited to no more than 50 nodes. We use a two-level local search heuristic to further enhance the results, pushing the limits to up to 300 nodes: a first local search step is directly integrated into the crossover operator; another one is applied independently during offspring generation.

Window-Based Multi-Objective Optimization for Dynamic Patient Scheduling with Problem-Specific Operators

The problem of patient admission scheduling (PAS) is a nondeterministic polynomial time (NP)-hard combinatorial optimization problem with numerous constraints. Researchers have divided the constraints of this problem into hard (i.e., feasible solution) and soft constraints (i.e., quality solution). The majority of research has dealt with PAS using integer linear programming (ILP) and single objective meta-heuristic searching-based approaches. ILP-based approaches carry high computational demand and the risk of non-feasibility for a large dataset. In a single objective optimization, there is a risk of local minima due to the non-convexity of the problem. In this article, we present the first pareto front-based optimization for PAS using set of meta-heuristic approaches. We selected four multi-objective optimization methods. Problem-specific operators were developed for each of them. Next, we compared them with single objective optimization approaches, namely, simulated annealing and particle swarm optimization. In addition, this article also deals with the dynamical aspect of this problem by comparing historical window-based decomposition with day decomposition, as has previously been proposed in the literature. An evaluation of the models proposed in the article and comparison with traditional models reveals the superiority of our proposed multi-objective optimization with window incorporation in terms of optimality.

Retail store layout optimization for maximum product visibility

It is well-established that increased product visibility to shoppers leads to higher sales for retailers. In this study, we propose an optimization methodology which assigns product categories and subcategories to store locations and sublocations to maximize the overall visibility of products to shoppers. The methodology is hierarchically developed to meet strategic and tactical layout planning needs of brick-and-mortar retailers. Layouts in both levels of planning are optimized considering eligibility requirements and complete set of shopper paths, thus, they successfully capture the unique shopping behaviour of consumers in a store’s region. The resulting mathematical optimization problem is recognized as a special instance of the well-known Quadratic Assignment Problem, which is considered computationally as one of the hardest optimization problems. We adopt a linearization technique and demonstrate via a real-world numerical example that our linearized optimization models substantially improve the store layout, hence, can be used in practical applications as a vital decision support model for store layout planning.

Cycle-tree guided attack of random K-core: Spin glass model and efficient message-passing algorithm

The K-core of a graph is the maximal subgraph within which each vertex is connected to at least K other vertices. It is a fundamental network concept for understanding threshold cascading processes with a discontinuous percolation transition. A minimum attack set contains the smallest number of vertices whose removal induces complete collapse of the K-core. Here we tackle this prototypical optimal initial-condition problem from the spin-glass perspective of cycle-tree maximum packing and propose a cycle-tree guided attack (CTGA) message-passing algorithm. The good performance and time efficiency of CTGA are verified on the regular random and Erd\"os-R\'enyi random graph ensembles. Our central idea of transforming a long-range correlated dynamical process to static structural patterns may also be instructive to other hard optimization and control problems.

ResLT: Residual Learning for Long-Tailed Recognition.

Deep learning algorithms face great challenges with long-tailed data distribution which, however, is quite a common case in real-world scenarios. Previous methods tackle the problem from either the aspect of input space (re-sampling classes with different frequencies) or loss space (re-weighting classes with different weights), suffering from heavy over-fitting to tail classes or hard optimization during training. To alleviate these issues, we propose a more fundamental perspective for long-tailed recognition, i.e., from the aspect of parameter space, and aims to preserve specific capacity for classes with low frequencies. From this perspective, the trivial solution utilizes different branches for the head, medium, tail classes respectively, and then sums their outputs as the final results is not feasible. Instead, we design the effective residual fusion mechanism - with one main branch optimized to recognize images from all classes, another two residual branches are gradually fused and optimized to enhance images from medium+tail classes and tail classes respectively. Then the branches are aggregated into final results by additive shortcuts. We test our method on several benchmarks, i.e., long-tailed version of CIFAR-10, CIFAR-100, Places, ImageNet, and iNaturalist 2018. Experimental results manifest the effectiveness of our method. Our code is available at https://github.com/jiequancui/ResLT.

Low Latency Aware Fog Nodes Placement in Internet of Things Service Infrastructure

Fog Computing extends storage and computation resources closer to end-devices. In several cases, the Internet of Things (IoT) applications that are time-sensitive require low response time. Thus, reducing the latency in IoT networks is one of the essential tasks. To this end, fog computing is developed with a motive for the data production and consumption to always be within proximity; therefore, the fog nodes must be placed at the edge of the network, which is near the end devices, such that the latency is minimized. The optimal location selection for fog node placement within a network out of a very large number of possibilities, such as minimize latency, is a challenging problem. So, it is a combinatorial optimization problem. Hard combinatorial optimization problems (NP-hard) involve huge discrete search spaces. The fog node placement problem is an NP-hard problem. NP-hard problems are often addressed by using heuristic methods and approximation algorithms. Combinatorial optimization problems can be viewed as searching for the best element of some set of discrete items; therefore in principle, any metaheuristic can be used to solve them. To resolve this, meta-heuristic-based methods is proposed. We apply the Simulated Annealing (SA), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) technique to design fog node placement algorithms. Genetic Algorithm is observed to give better solutions. Since Genetic Algorithm may get stuck in local optima, Hybrid Genetic Algorithm, and Simulated Annealing (GA-SA), Hybrid Genetic Algorithm and Particle Swarm Optimization (GA-PSO) were compared with GA. By extensive simulations, it is observed that hybrid GA-SA-based for node placement algorithm outperforms other baseline algorithms in terms of response time for the IoT applications.

How to Assess Different Algorithms Using Intuitionistic Fuzzy Logic

Intuitionistic fuzzy logic is the main tool in the recently developed step-wise “cross-evaluation” procedure that aims at the assessment of different optimization algorithms. In this investigation, the procedure previously applied to compare the effectiveness of two or three algorithms has been significantly upgraded to evaluate the performance of a set of four algorithms. For the first time, the procedure applied here has been tested in the evaluation of the effectiveness of genetic algorithms (GAs), which are proven as very promising and successful optimization techniques for solving hard non-linear optimization tasks. As a case study exemplified with the parameter identification of a S. cerevisiae fed-batch fermentation process model, the cross-evaluation procedure has been executed to compare four different types of GAs, and more specifically, multi-population genetic algorithms (MGAs), which differ in the order of application of the three genetic operators: Selection, crossover and mutation. The results obtained from the implementation of the upgraded intuitionistic fuzzy logic-based procedure for MGA performance assessment have been analyzed, and the standard MGA has been outlined as the fastest and most reliable one among the four investigated algorithms.

Improvement and Optimization of the Hardness for the Aluminum Metal Matrix Composite Using Eggshell

This research examines the effect of size and weight fraction of the eggshell particles added on the hardening of the Al metal reinforced by eggshell Particles using the program MINITAB 16. Powder technology method has been used to produce a composite material, The composite material was obtained by adding different weight fractions (0,2, 6,10 and 12 wt. %) and the size (100,240,350,510and 670um) of eggshell particles to aluminum powder. It samples was composed by using pressing of single action then accompanied directly by process of sintering at 474°C under the inert gas effect conditions. it was concluded that the best hardening of (72.6971 Mpa) can be obtained when the variables were (Xi = 12 wt%) and (X2 =100um). The lights value of Xi and X2, obtained using the programs, was used in practice giving hardening (74Mpa) which it nearly similar to that obtained by program. Also results show that the variables weight fraction (X1) &amp; Particle Size (X2) have a significant effect on hardening. Moreover, the hardening increases with increasing weight fraction. While hardening value decreases gradient with down eggshell particle size.

Investigating Process Parameters of Abrasive Water Jet Machine using EN31

ABSTRACT This document explains the machining of EN31 steel using an Abrasive Water Jet Machine. The work is aimed at investigating the role of Pressure, Abrasive Mass Flow Rate, Traverse Speed, and Standoff Distance on; Machining Time, Surface Roughness, and Hardness, followed by the optimization of these responses. Taguchi approach and Analysis of Variance are practiced to investigate the effecting features of the process parameters, regarding the responses. A regression model is also developed for efficient study. The experimental optimal data and the responses are calculated through the machining process, while the predicted optimal data is calculated through the various factorial effects plots for signal to noise ratios (SNRs). At the predicted optimal data, the responses are calculated. The validation test confirms the predicted responses and hence the acceptability of the specified model. The experiment reveals the various levels of the particular parameters in the optimization of Machining Time, Surface Roughness, and Hardness. Also, the major and minor contribution of the different parameters is investigated.

DOE-based experimental investigation and optimization of hardness and corrosion rate for Cu-x%Al&lt;sub&gt;2&lt;/sub&gt;O&lt;sub&gt;3&lt;/sub&gt; as processed by powder metallurgy

<abstract> Copper alumina (Cu-x%Al₂O₃) was prepared from micro-sized powder particles through the powder metallurgy (PM) process. Experimental runs were designed according to the design of experiments (DOE) methodology to investigate the effect of alumina concentration and sintering time on the hardness and corrosion rate of the powdered composite. Four sintering times were tested (1, 2, 3, and 4 h) in combination with different levels of alumina concentration (1 wt%, 3 wt%, 6 wt%, and 10 wt%). Given the narrow range of the sintering temperature for copper alumina, this temperature was fixed at 900 ℃, whereas the compaction pressure was set at 300 MPa. Collected data were analysed based on the analysis of variance (ANOVA) approach and complemented with microscopic analysis. Optimization techniques were used to identify the optimal settings of sintering time and alumina concentration for the PM composite. </abstract>

On modelling hard combinatorial optimisation problems as linear programs: refutations of the 'unconditional impossibility' claims

On modelling hard combinatorial optimisation problems as linear programs: refutations of the 'unconditional impossibility' claims

Symmetrical Synthesis for Deep Metric Learning

Deep metric learning aims to learn embeddings that contain semantic similarity information among data points. To learn better embeddings, methods to generate synthetic hard samples have been proposed. Existing methods of synthetic hard sample generation are adopting autoencoders or generative adversarial networks, but this leads to more hyper-parameters, harder optimization, and slower training speed. In this paper, we address these problems by proposing a novel method of synthetic hard sample generation called symmetrical synthesis. Given two original feature points from the same class, the proposed method firstly generates synthetic points with each other as an axis of symmetry. Secondly, it performs hard negative pair mining within the original and synthetic points to select a more informative negative pair for computing the metric learning loss. Our proposed method is hyper-parameter free and plug-and-play for existing metric learning losses without network modification. We demonstrate the superiority of our proposed method over existing methods for a variety of loss functions on clustering and image retrieval tasks.

Optimization of hardness and wear parameters of Al-Cu-Si alloy using design of experiments

Abstract Aluminium-Copper-Silicon (Al-Cu-Si) alloy of composition Al-1.5Cu-0.5Si alloy was successfully fabricated by sand casting process. Homogenization treatment was done at different temperatures to achieve uniform distribution of elements. The hardness was measured in different regions (outer, middle and inner) and the peak hardness of the alloy was obtained at 425 °C homogenization temperature in the outer region because of low solidification time at that region. L9 Orthogonal array was used to conduct the experiments and optimize the process parameters in the three regions for Specific Wear Rate (SWR) and the influencing parameters such as load, speed, homogenization temperature and sliding time were analyzed using Taguchi techniques. From the Analysis of variance (ANOVA), it was observed that for outer region, the sliding speed has the major effect on SWR followed by load, sliding time and homogenization temperature. In case of middle and inner regions, the influence of load on SWR was significant followed by sliding speed. The confirmation test was carried with obtained optimal process parameter. The surfaces wear mechanism behind Al-1.5Cu-0.5Si was analysed using scanning electron microscope (SEM) image

Preliminary exploration of hard X-ray coherent diffraction imaging method at SSRF

Coherent X-ray diffraction imaging (CDI) method is a powerful X-ray imaging technique with high resolution up to nanometer scale. Most of the synchrotron radiation facilities and free electron laser facilities are equipped with this state-of-the-art imaging technique and have made many outstanding achievements in multiple scientific areas. Up to now, although scanning CDI (ptychography) method based on a soft X-ray source has been opened to users, the hard X-ray CDI experimental platform has not been built at Shanghai Synchrotron Radiation Facility (SSRF) which can research some relatively thick specimens and easily extend to three-dimensional imaging. As some new beamlines with undulator source were put into operation recently, it is possible and feasible to build up the CDI experimental platform with hard X-ray. In this article, we report the hard X-ray CDI experimental platform development process and preliminary experimental results of coherent diffraction pattern and image reconstruction at SSRF. Based on the operating BL19U2 biological small-angle X-ray scattering (SAXS) beamline at SSRF, the hard X-ray coherent beam is obtained through effective optical path designation at 12 keV and 13.5 keV. The hard X-ray optimization includes tuning several slits, double crystal monochromator (DCM), horizontal deflection mirror, focusing mirror system and pinhole, etc. Furthermore, hard X-ray CDI experiments are conducted. The spatial coherent length of the incident beam is also measured from the pinhole diffraction pattern. This platform can provide both conventional mode and scanning mode (ptychography) for the coherent diffraction imaging method, and the correct image reconstruction from the experimental diffraction patterns proves that the platform has the experimental capability for hard X-ray CDI. In the conventional forward scattering CDI mode, coherent diffraction patterns of pinhole are collected and used to analyse the coherence property of the optimized X-ray beam. The structure of pinhole is also reconstructed from the diffraction pattern. In the scanning CDI mode, a zone plate is used as a sample. The central area of zone plate is reconstructed correctly. About 90 nm/pixel resolution of reconstruction is achieved which is extremely dependent on the X-ray flux density from the undulator source emission. Hard X-ray CDI experimental platform based on the synchrotron radiation facility is first built in China. It will provide effective software and hardware supporting for the development and application of hard X-ray CDI experiments in China in the future.

Capacity Allocation of Game Tickets Using Dynamic Pricing

This study examines a pricing approach that is applicable in the field of online ticket sales for game tickets. The mathematical principle of dynamic programing is combined with empirical data analysis to determine demand functions for university football game tickets. Based on the calculated demand functions, the application of DP strategies is found to generate more revenues than a fixed price strategy. The other important result is the capacity distribution of tickets according to the football game intensity. Prior studies have shown that it is sometimes more profitable or football clubs to allocate a share of tickets to a retailer and earn a commission based on the sales, rather than selling the entire capacity of tickets by itself. This paper finds that in a high intensity game, where the demand is generally high, it is optimal for the club to sell all tickets by itself. Whereas, for less popular games, where there is considerable fluctuation in demand, the capacity allocation problem for maximized revenues from ticket sales, becomes a harder optimization challenge for the club. According to DP optimization, when the demand for tickets is relatively low, it is optimal for the club to retain 20–40% of the tickets and the rest of the capacity should be sold to online retailers. In the real world, this pricing technique has been used by football clubs and thus the secondary market online retailers like Ticketmaster and Vivid Seats have become popular in the last decade.

Solving Travelling Salesman Problem Using Ant Systems: A Programmer’s Approach

For an engineer, optimization is the process of determining the best design. In this work, a computer-based approach to a design optimization is investigated. This enables us to evaluate many design combinations that cannot be done manually. Travelling salesman problem (TSP) is leading problem among NP hard combinatorial optimization problems. It is widely used in many engineering applications. Ant system is a heuristic approach used to solve combinatorial optimization problems. In this work, a MATLAB program to solve TSP using Ant Search method is developed and is explained through the perspective of a programmer. Mathematical model of the problem is presented along with a detailed flowchart to describe the algorithm. The proposed program is described in accordance with the flow chart. Results obtained for two cases are summarized using graphs and CPU times. Finally a brief study on tuning of parameters of the method is also presented.