Real Design Problems Research Articles

CMGWO: Grey wolf optimizer for fusion cell-like P systems

The grey wolf optimizer is a widely used parametric optimization algorithm. It is affected by the structure and rank of grey wolves and is prone to falling into the local optimum. In this study, we propose a grey wolf optimizer for fusion cell-like P systems. Cell-like P systems can parallelize computation and communicate from cell membrane to cell membrane, which can help the grey wolf optimizer jump out of the local optimum. Design new convergence factors and use different convergence factors in other cell membranes to balance the overall exploration and utilization capabilities of the algorithm. At the same time, dynamic weights are introduced to accelerate the convergence speed of the algorithm. Experiments are performed on 24 test functions to verify their global optimization performance. Meanwhile, a support vector machine model optimized by the grey wolf optimizer for fusion cell-like P systems has been developed and tested on six benchmark datasets. Finally, the optimizing ability of grey wolf optimizer for fusion cell-like P systems on constrained optimization problems is verified on three real engineering design problems. Compared with other algorithms, grey wolf optimizer for fusion cell-like P systems obtains higher accuracy and faster convergence speed on the test function, and at the same time, it can find a better parameter set stably for the optimization of support vector machine parameters, in addition to being more competitive on constrained engineering design problems. The results show that grey wolf optimizer for fusion cell-like P systems improves the searching ability of the population, has a better ability to jump out of the local optimum, has a faster convergence speed, and has better stability.

Using senior undergraduate students as coaches in an introductory chemical engineering design course

Senior undergraduate students were hired as Teaching Assistants (TAs) in CHBE 221, an introductory bioprocess design course offered in the Department of Chemical and Biological Engineering at University of British Columbia, and the roles were recast from traditional course support, to being coaches and mentors to a small number of teams working on term-spanning design projects. This was done as a means to provide more personalized support to students working on realistic design problems early in their training, leverage the benefits of peer learning strategies, and provide professional and leadership development opportunities to the TAs ahead of graduation. At the end of the course in which this was implemented, surveys were sent to students asking them to comment on their experience with this mentorship model, and the TAs of the course were interviewed to better understand theirs. This paper presents the results of these surveys and interviews, highlights the benefits and challenges of this approach, and provides recommendations for future implementations of this student support model

PHYSICS-INFORMED NEURAL NETWORKS FOR MODELING OF 3D FLOW THERMAL PROBLEMS WITH SPARSE DOMAIN DATA

Successfully training physics-informed neural networks (PINNs) for highly nonlinear partial differential equations (PDEs) on complex 3D domains remains a challenging task. In this paper, PINNs are employed to solve the 3D incompressible Navier-Stokes equations at moderate to high Reynolds numbers for complex geometries. The presented method utilizes very sparsely distributed solution data in the domain. A detailed investigation of the effect of the amount of supplied data and the PDE-based regularizers is presented. Additionally, a hybrid data-PINNs approach is used to generate a surrogate model of a realistic flow thermal electronics design problem. This surrogate model provides near real-time sampling and was found to outperform standard data-driven neural networks (NNs) when tested on unseen query points. The findings of the paper show how PINNs can be effective when used in conjunction with sparse data for solving 3D nonlinear PDEs or for surrogate modeling of design spaces governed by them.

Design for Additive Manufacturing Education of Process Engineering Students on an Industrial Challenge

As the use of additive manufacturing (AM) is increasing in many sectors, the need of design education for AM has gained importance. Solving the challenges of an application with the advantages of AM requires competencies in selecting a suitable AM process, identifying design opportunities and designing for AM based on an extensive knowledge in the application domain. Hence, it is necessary for all branches engineering to develop AM skills and competencies. This paper describes a lecture and practical to develop AM design competencies in chemical process engineering master students. The course teaches designing complex AM parts based on the need of an application by combining the opportunistic and restrictive aspects AM design with the domain-specific functional requirements. The design education benefits from AM’s advantages in rapid prototyping to provide feedback within the time frame of a course. In order to fulfill the above requirements, an application-oriented design course from requirements to testing, involving a simplified real industrial design problem, has been prepared. Students obtain design requirements at the beginning of the practical and have to develop a device using the advantages of AM. After completing CAD design, the designs of the students are produced with the laser based powder-bed fusion of metal (PBF-LB/M) and bath-based photopolymerisation of polymers, curing using ultraviolet light (VPP-UVL/P) methods. The manufactured designs are tested on the test bench provided by the institute. A case study of a hybrid manufactured reactor with internal condensation supports the conclusion of this paper.

A step-size follow-the-leader optimization algorithm with an improved step parameters

Follow the leader (FTL) algorithm is a newly developed optimization algorithm inspired by a sheep’s movement within a flock. FTL has been successfully implemented to solve power prediction problems. However, the probability of falling in local optima is high due to randomness in the step parameter. This paper proposes a step-size follow-the-leader (SFTL) algorithm with decreasing and increasing step-size combinations. The improved step parameter tunes the search space by generating a new solution to improve the accuracy and convergence rate of the FTL algorithm. Four different FTL variants have been presented in this paper to show the impact of the dynamic step-size parameter. FTL improvement is verified by testing SFTL over thirty-two fixed unimodal, unimodal, fixed multimodal, and multimodal benchmark functions. The computational results indicate that SFTL significantly improves the basic FTL algorithm and converges early compared to other algorithms. SFTL has also been tested on five real engineering design problems and obtained results show that SFTL outperformed FTL and other popular optimization algorithms.

Robust optimization of the design of monopropellant propulsion control systems using an advanced teaching-learning-based optimization method

This research proposes a novel approach for the robust optimization of the design of hydrogen peroxide propulsion control systems using the efficient and advanced Teaching-Learning-Based Optimization (TLBO) method. This study adopts a robust design optimization (RDO) formulation that considers both epistemic and aleatory uncertainties, including sparse points and interval data, and uses the Johnson distribution family for uncertainty representation. The maximum likelihood estimation method is applied to determine the distribution parameters, also considering interval data with a nested optimization technique. A novel advanced TLBO method with high accuracy and convergence rate is employed to optimization of this robust design approach. The method’s originality and advancement come from two categories of modifications to the original framework: the structure of the teaching and learning phases and the initialization and search approach. The efficacy and applicability of the proposed Ad-TLBO, respectively, were evaluated using benchmark problems from the CEC2020 competition and three real-world engineering problems, with a comparison to some recently published and the CEC competition’s top-ranked algorithms. The results and statistical analyses of the Quade test, Wilcoxon signed-rank test, and Friedman test show that the proposed Ad-TLBO method outperforms the other algorithms. The proposed optimization method is eventually applied to design a monopropellant propulsion system as the control actuator of a satellite orbital transfer system. It is found that the proposed advanced TLBO is effective in handling uncertainty in real design problems and improves both the convergence rate and accuracy of the optimization process.

VIRTUALLY HOSTED HACKATHONS FOR DESIGN RESEARCH: LESSONS LEARNED FROM THE INTERNATIONAL DESIGN ENGINEERING ANNUAL (IDEA) CHALLENGE 2022

AbstractThe International Design Engineering Annual (IDEA) Challenge is a virtually hosted hackathon for Engineering Design researchers with aims of: i) generating open access datasets; ii) fostering community between researchers; and, iii) applying great design minds to develop solutions to real design problems. This paper presents the 2022 IDEA challenge and elements of the captured dataset with the aim of providing insights into prototyping behaviours at virtually hosted hackathons, comparing it with the 2021 challenge dataset and providing reflections and learnings from two years of running the challenge. The dataset is shown to provide valuable insights into how designers spend their time at hackathon events and how, why and when prototypes are used during their design processes. The dataset also corroborates the findings from the 2021 dataset, demonstrating the complementarity of physical and sketch prototypes. With this paper, we also invite the wider community to contribute to the IDEA Challenge in future years, either as participants or in using the platform to run their own design studies.

A novel improved chef-based optimization algorithm with Gaussian random walk-based diffusion process for global optimization and engineering problems

The chef-based optimization algorithm (CBOA) is a human-based method inspired by the relationship between culinary students and chef instructors. The original CBOA does not have a process that separately controls population generation and updating. For this reason, in this study, a novel improved CBOA with diffusion process (DP) algorithm with the Gaussian walk (CBOADP) is proposed. In this way, it is aimed to transfer promising members to the next iteration. In the original CBOA algorithm, the selection of the chefs is performed randomly. The oppositional-based learning (OBL) approach aims to find the current best chef by updating the chef from the chef candidates selected from opposite regions. Because the selection of the chef plays an important role in the selection of students. To develop the method with optimal parameters, parameter analysis is performed as the DP’s walk and maximum diffusion parameters. The new method’s performance is evaluated through the CEC2019 and CEC2022 test suites and traditional and advanced methods. Results for real design problems are also reported. All results are supported by the Wilcoxon sign and Friedman rank statistical tests. As a result, in general, the best convergence has been achieved with CBOADP in about 75% of the benchmark test data, and statistically, it has first rank among other methods on average. In addition, CBOADP reached the best fitness value in five of the six compared engineering problems. Accordingly, it is seen that DP and OBL approaches contribute positively to the performance of CBOA and show a successful convergence trend among other compared methods.

Chaotic enhanced teaching-based differential evolution algorithm applied to discrete truss optimization

In this paper, we develop a discrete chaotic enhanced hybrid teaching-based differential evolution algorithm (CETDE) that can deal with realistic yet challenging constrained discrete truss design problems more efficiently. CETDE adaptively executes multiple mutation operators to quickly search for the solution, but at the same time, premature convergence, a common issue in discrete optimization, is also prevented by embedding the chaotic logistic rule into the mutations. An improved logical strategy that only allows potential mutants to go through structural analysis is introduced to reduce computation. An enhanced chaotic local search strategy and a one-component change technique are also developed for further efficiency. Besides, CETDE simultaneously keeps continuous and discrete population throughout the evolution, which is distinguishable from existing discrete optimizers, to preserve the stochastic nature. Compared to other state-of-the-art performers in the literature, CETDE illustrates both higher efficiency and solution optimality.

Ọdịgbo Metaheuristic Optimization Algorithm for Computation of Real-Parameters and Engineering Design Optimization

This paper proposes a new population-based global optimization algorithm, Ọdịgbo Metaheuristic Optimization Algorithm–ỌMOA, for solving complex bounded-constraint/single objective real-parameter problems found in most engineering and scientific applications. It’s inspired by the human socio-cultural informal discipleship learning pattern inherent in the behavior of the Ndịgbo peoples; the subject – primary (Nwa-ahịa), in mercantile cycle grows to a secondary (Mazi) owing to the intuitive stratagem (dialect - Ịgba) embedded in an aged-long cultural model “Ịgba-ọsọ-ahịa” (meaning, strategic marketing skills, and practice). The model mimics the search routine for satisfying a customer’s need in the market, built into exploration and exploitation applied in the mathematical model. About 30 complex classical unconstrained functions are tested, comparing results with that of five similar state-of-the-art algorithms. Also, 29 CEC-2017 single objective real constraint benchmark serious dimensional problems were simulated and compared against the winners of that competition. Validation includes statistical (t-test, p-value) comparison and for 50 Dimension constraint problems as ỌMOA demonstrated superior performance. TCS (9.18%), WBP (6.3%), PVDP (601%), RGP (319%), RBP (760%), GTCD (202%), HIMMELBLAU (4%), and CDP (88.12%) are the improvements made on 8 CEC-2020 engineering real design problems against the former best performances; OMOA is simple to implement, replicate and applicable across domains. Also, some new, improved optimum was obtained in Shubert and Schaffer 4 function compared to the global optimums.

Termite life cycle optimizer

This paper introduces a novel bio-inspired meta-heuristic optimization algorithm, named termite life cycle optimizer (TLCO), which is based on both the life cycle of a termite colony and the modulation of movement strategies used by many animal species in nature. Termite colonies are comprised of three distinct castes: the workers, the soldiers and the reproductive termites. Each caste undertakes a set of specific tasks that ensure the growth and survival of the colony. TLCO mimics the activities of these three castes that are implemented in a mathematical model. The model is then used to find the global optimum in classic optimization problems. First, the behaviors of the workers, soldiers and reproductive termites are used to simulate a balance between the tasks of exploration and exploitation. Second, the initial population securely records the information obtained at each iteration and transmits it to workers and soldiers at the next iteration. This process is repeated until the global optimum is found with the smallest error. Besides, a new proposed function combined with Lévy flight is used to modulate the movement of termites that increases its flexibility. Thus, TLCO can cover both long distances during the first iterations to improve the convergence rate and shorter distances during the last iterations to enhance the level of accuracy. We then compare the performances of TLCO with other well-known nature-inspired algorithms using 23 classical benchmark functions, CEC2005 benchmark functions, and five real engineering design problems. The results demonstrate the effectiveness and reliability of TLCO in solving these optimization problems. Source codes of TLCO is publicly available at http://goldensolutionrs.com/termite-life-cycle-optimizer.html.

Oppositional Slime Mould Algorithm: Development and application for designing demand side management controller

In recent years, application of several metaheuristic optimization algorithms have been reported. Their flexibility and adaptive structure make them compatible for real engineering design problems. The ability of the optimization algorithm for solving any design problem depends upon the balance between exploration and exploitation phase of the algorithm. These phases are determined by inherent mathematical equations. For making any algorithm compatible for engineering design tasks, certain modifications are required. In this work, we apply two modifications namely (opposition theory and sine cosine based position update mechanism for enhancing exploration and exploitation virtues of recently introduced Slime Mould Algorithm (SMA). Further, the suitability of designing a Demand Side Management Controller by proposed algorithm is explored with the help of standard mathematical analysis and bench marking procedures. Comparison of DSM controllers based on different algorithms has also been evaluated. The results reveal that proposed modifications enhances the optimization capability of SMA.

Constrained robust Bayesian optimization of expensive noisy black‐box functions with guaranteed regret bounds

AbstractMany real‐world design problems involve optimization of expensive black‐box functions. Bayesian optimization (BO) is a promising approach for solving such challenging problems using probabilistic surrogate models to systematically tradeoff between exploitation and exploration of the design space. Although BO is often applied to unconstrained problems, it has recently been extended to the constrained setting. Current constrained BO methods, however, cannot identify solutions that are robust to unavoidable uncertainties. In this article, we propose a robust constrained BO method, constrained adversarially robust Bayesian optimization (CARBO), that addresses this challenge by jointly modeling the effect of the design variables and uncertainties on the unknown functions. Using exact penalty functions, we establish a bound on the number of CARBO iterations required to find a near‐global robust solution and provide a rigorous proof of convergence. The advantages of CARBO are demonstrated on two case studies including a non‐convex benchmark problem and a realistic bubble column reactor design problem.

A new hybrid method based on Aquila optimizer and tangent search algorithm for global optimization.

Since no single algorithm can provide the optimal solutions for all problems, new metaheuristic methods are always being proposed or developed by combining current algorithms or creating adaptable versions. Metaheuristic methods should have a balanced exploitation and exploration stages. One of these two talents may be sufficient in some metaheuristic methods, while the other may be insufficient. By integrating the strengths of the two algorithms and hybridizing them, a more efficient algorithm can be formed. In this paper, the Aquila optimizer-tangent search algorithm (AO-TSA) is proposed as a new hybrid approach that uses the intensification stage of the tangent search algorithm (TSA) instead of the limited exploration stage to improve the Aquila optimizer’s exploitation capabilities (AO). In addition, the local minimum escape stage of TSA is applied in AO-TSA to avoid the local minimum stagnation problem. The performance of AO-TSA is compared with other current metaheuristic algorithms using a total of twenty-one benchmark functions consisting of six unimodal, six multimodal, six fixed-dimension multimodal, and three modern CEC 2019 benchmark functions according to different metrics. Furthermore, two real engineering design problems are also used for performance comparison. Sensitivity analysis and statistical test analysis are also performed. Experimental results show that hybrid AO-TSA gives promising results and seems an effective method for global solution search and optimization problems.

High-Fidelity Surrogate Based Multi-Objective Optimization Algorithm

The employment of conventional optimization procedures that must be repeatedly invoked during the optimization process in real-world engineering applications is hindered despite significant gains in computing power by computationally expensive models. As a result, surrogate models that require far less time and resources to analyze are used in place of these time-consuming analyses. In multi-objective optimization (MOO) problems involving pricey analysis and simulation techniques such as multi-physics modeling and simulation, finite element analysis (FEA), and computational fluid dynamics (CFD), surrogate models are found to be a promising endeavor, particularly for the optimization of complex engineering design problems involving black box functions. In order to reduce the expense of fitness function evaluations and locate the Pareto frontier for MOO problems, the automated multiobjective surrogate based Pareto finder MOO algorithm (AMSP) is proposed. Utilizing data samples taken from the feasible design region, the algorithm creates three surrogate models. The algorithm repeats the process of sampling and updating the Pareto set, by assigning weighting factors to those surrogates in accordance with the values of the root mean squared error, until a Pareto frontier is discovered. AMSP was successfully employed to identify the Pareto set and the Pareto border. Utilizing multi-objective benchmark test functions and engineering design examples such airfoil shape geometry of wind turbine, the unique approach was put to the test. The cost of computing the Pareto optima for test functions and real engineering design problem is reduced, and promising results were obtained.

On Hierarchical Applications of Finite Element Methods and Classical Applied Mechanics Approaches for Complex Structures

In this paper, hierarchical studies based on both analytical and computational approaches have been employed to validate and improve current structural designs of mud holding tanks for hydraulic fracturing in petroleum industries. Through detailed comparisons at different levels of approximations for both real world complex structural design problem and a simplified stiffened plate with extreme geometrical aspect ratios, proper design procedures have been reiterated and validated. It is highly recommended that low-dimensional structures with emphasis on classical mechanics principles should be considered before full-fledged three-dimensional computational analyses. Both Castigliano’s Theorems and Galerkin methods are utilized along with simulation packages such as ADINA and Solidworks. The hierarchical approximation procedures reemphasize important subjects such as planes and shells, strength of materials, structural mechanics, asymptotic and perturbation analysis no longer required in some engineering programs, yet often necessary for the study of complex three-dimensional structural systems.

GGWO: Gaze cues learning-based grey wolf optimizer and its applications for solving engineering problems

In this article, an improved variant of the grey wolf optimizer (GWO) named gaze cues learning-based grey wolf optimizer (GGWO) is proposed. The main intentions are to reduce the existing high selective pressure and low diversification of the GWO algorithm, which results in premature convergence, local optima trapping, and stagnation problems. The GGWO algorithm benefits from two new search strategies: neighbor gaze cues learning (NGCL) and random gaze cues learning (RGCL) inspired by the gaze cueing behavior in wolves. The NGCL strategy enhances the exploitation ability and local optima avoidance. The RGCL, however, boosts the population diversity and balance between exploration and exploitation. The cooperation among three search strategies GWO, NGCL, and RGCL, improves diversification, exploration, and exploitation. The GGWO algorithm performance was evaluated by conducting CEC'18 test functions. Furthermore, the results of GGWO were compared with nine metaheuristic algorithms KH, iwPSO, WOA, GWO, GWO-EPD, HGWOSCA, EEGWO, BOA, and VAGWO. Moreover, the experimental results were statistically analyzed by the Wilcoxon signed-rank and Friedman tests. Additionally, four real engineering design problems and two problems of optimal power flow (OPF) for the IEEE 30-bus and IEEE 118-bus are optimized to verify the applicability of the GGWO in practice. The results show that the GGWO algorithm has been able to provide competitive and superior results to the compared algorithms, and it is capable of solving engineering problems.

Bayesian optimization for goal-oriented multi-objective inverse material design

SummaryBayesian optimization (BO) can accelerate material design requiring time-consuming experiments. However, although most material designs require tuning of multiple properties, the efficiency of multi-objective (MO) BO in time-consuming experimental material design remains unclear, due to the complexity of handling multiple objectives. This study introduces MO BO method that efficiently achieves predefined goals and shows that by focusing on achieving the goals, BO can efficiently accelerate realistic MO design problems with small efforts. Benchmarks showed that the proposed BO method dramatically reduced the number of experiments needed to achieve goals relative to a baseline method. Virtual MO inverse design experiments with realistic material design problems were also performed, during which the proposed method could achieve goals within only around ten experiments in average and showed over 1000-fold acceleration relative to the random sampling for the most difficult case. The introduction of goal-oriented BO will precede real-world application of BO.

Comparison of metaheuristic optimization algorithms for solving constrained mechanical design optimization problems

Determining the solution for real mechanical design problems is a challenging task when using the newly developed and efficient swarm intelligence algorithms. There are so many difficulties to be addressed, including but not limited to mixed decision variables, diverse constraints, inherent errors, conflicting objectives, and numerous locally optimal solutions. This work analyzes the behavior of nine metaheuristic algorithms, namely, salp swarm algorithm (SSA), multi-verse optimizer (MVO), moth-flame optimizer (MFO), atom search optimization (ASO), ecogeography-based optimization (EBO), queuing search algorithm (QSA), equilibrium optimizer (EO), evolutionary strategy (ES) and hybrid self-adaptive orthogonal genetic algorithm (HSOGA). The efficiency of these algorithms is evaluated on eight mechanical design problems using the solution quality and convergence analysis, which verifies the wide applicability of these algorithms to real-world application problems.

A comparative analysis of the queuing search algorithm, the sine-cosine algorithm, the ant lion algorithm to determine the optimal weight design problem of a spur gear drive system

Abstract Metaheuristic optimization algorithms have gained relevance and have effectively been investigated for solving complex real design problems in diverse fields of science and engineering. In this paper, a recent meta-heuristic approach inspired by human social concepts, namely the queuing search algorithm (QSA), is implemented for the first time to optimize the main parameters of the spur gear, in particular, to minimize the weight of a single-stage spur gear. The effectiveness of the algorithm introduced is examined in two steps. First, the algorithm used is compared with descriptions in previous studies and indicates that the final results obtained by QSA lead to a reduction in gear weight by 7.5 %. Furthermore, the outcomes obtained are compared with those for the other five algorithms. The results reveal that the QSA outperforms the techniques with which it is compared such as the sine-cosine optimization algorithm, the ant lion optimization algorithm, the interior search algorithm, the teaching-learning-based algorithm, and the jaya algorithm in terms of robustness, success rate, and convergence capability.