scholarly journals An Exploratory Landscape Analysis-Based Benchmark Suite

Algorithms ◽  
2021 ◽  
Vol 14 (3) ◽  
pp. 78
Author(s):  
Ryan Dieter Lang ◽  
Andries Petrus Engelbrecht
Keyword(s):  
Sample Size ◽  
Landscape Analysis ◽  
Benchmark Problems ◽  
Objective Functions ◽  
Algorithm Selection ◽  
Problem Space ◽  
Constrained Problem ◽  
Benchmark Suite ◽  
Selection Framework ◽  
Crucial Part

The choice of which objective functions, or benchmark problems, should be used to test an optimization algorithm is a crucial part of the algorithm selection framework. Benchmark suites that are often used in the literature have been shown to exhibit poor coverage of the problem space. Exploratory landscape analysis can be used to quantify characteristics of objective functions. However, exploratory landscape analysis measures are based on samples of the objective function, and there is a lack of work on the appropriate choice of sample size needed to produce reliable measures. This study presents an approach to determine the minimum sample size needed to obtain robust exploratory landscape analysis measures. Based on reliable exploratory landscape analysis measures, a self-organizing feature map is used to cluster a comprehensive set of benchmark functions. From this, a benchmark suite that has better coverage of the single-objective, boundary-constrained problem space is proposed.

scholarly journals An evolutionary-based optimization algorithm for truss sizing design

2016 ◽  
Vol 38 (4) ◽  
pp. 307-317
Author(s):  
Pham Hoang Anh
Keyword(s):  
Local Search ◽  
Objective Function ◽  
Optimization Algorithm ◽  
Optimization Problems ◽  
Optimal Solution ◽  
The Other ◽  
Truss Structures ◽  
Benchmark Problems ◽  
Discrete Variables ◽  
Objective Functions

In this paper, the optimal sizing of truss structures is solved using a novel evolutionary-based optimization algorithm. The efficiency of the proposed method lies in the combination of global search and local search, in which the global move is applied for a set of random solutions whereas the local move is performed on the other solutions in the search population. Three truss sizing benchmark problems with discrete variables are used to examine the performance of the proposed algorithm. Objective functions of the optimization problems are minimum weights of the whole truss structures and constraints are stress in members and displacement at nodes. Here, the constraints and objective function are treated separately so that both function and constraint evaluations can be saved. The results show that the new algorithm can find optimal solution effectively and it is competitive with some recent metaheuristic algorithms in terms of number of structural analyses required.

An Evolutionary Algorithm for Black-Box Chance-Constrained Function Optimization

2013 ◽  
Vol 17 (2) ◽  
pp. 272-282
Author(s):  
Kazuyuki Masutomi ◽  
◽  
Yuichi Nagata ◽  
Isao Ono ◽  
◽  
...  
Keyword(s):  
Objective Function ◽  
Evolutionary Algorithm ◽  
Black Box ◽  
Function Optimization ◽  
Benchmark Problems ◽  
Objective Functions ◽  
Neural Network Controller ◽  
Objective Function Values ◽  
The Individual ◽  
Constrained Function

This paper presents an evolutionary algorithm for Black-Box Chance-Constrained Function Optimization (BBCCFO). BBCCFO is to minimize the expectation of the objective function under the constraints that the feasibility probability is higher than a userdefined constant in uncertain environments not given the mathematical expressions of objective functions and constraints explicitly. In BBCCFO, only objective function values of solutions and their feasibilities are available because the algebra expressions of objective functions and constraints cannot be used. In approaches to BBCCFO, a method based on an evolutionary algorithm proposed by Loughlin and Ranjithan shows relatively good performance in a realworld application, but this conventional method has a problem in that it requires many samples to obtain a good solution because it estimates the expectation of the objective function and the feasibility probability of an individual by sampling the individual plural times. In this paper, we propose a new evolutionary algorithm that estimates the expectation of the objective function and the feasibility probability of an individual by using the other individuals in the neighborhood of the individual. We show the effectiveness of the proposed method through experiments both in benchmark problems and in the problem of a inverted pendulum balancing with a neural network controller.

scholarly journals Systematic review of basket trials, umbrella trials, and platform trials: a landscape analysis of master protocols

Trials ◽  
2019 ◽  
Vol 20 (1) ◽  
Author(s):  
Jay J. H Park ◽  
Ellie Siden ◽  
Michael J. Zoratti ◽  
Louis Dron ◽  
Ofir Harari ◽  
...  
Keyword(s):  
Sample Size ◽  
English Language ◽  
Median Number ◽  
Landscape Analysis ◽  
Phase Iii ◽  
Experimental Drugs ◽  
Study Duration ◽  
Language Studies ◽  
Multiple Treatments ◽  
Basket Trials

Abstract Background Master protocols, classified as basket trials, umbrella trials, and platform trials, are novel designs that investigate multiple hypotheses through concurrent sub-studies (e.g., multiple treatments or populations or that allow adding/removing arms during the trial), offering enhanced efficiency and a more ethical approach to trial evaluation. Despite the many advantages of these designs, they are infrequently used. Methods We conducted a landscape analysis of master protocols using a systematic literature search to determine what trials have been conducted and proposed for an overall goal of improving the literacy in this emerging concept. On July 8, 2019, English-language studies were identified from MEDLINE, EMBASE, and CENTRAL databases and hand searches of published reviews and registries. Results We identified 83 master protocols (49 basket, 18 umbrella, and 16 platform trials). The number of master protocols has increased rapidly over the last five years. Most have been conducted in the US (n = 44/83) and investigated experimental drugs (n = 82/83) in the field of oncology (n = 76/83). The majority of basket trials were exploratory (i.e., phase I/II; n = 47/49) and not randomized (n = 44/49), and more than half (n = 28/48) investigated only a single intervention. The median sample size of basket trials was 205 participants (interquartile range, Q3-Q1 [IQR]: 500–90 = 410), and the median study duration was 22.3 (IQR: 74.1–42.9 = 31.1) months. Similar to basket trials, most umbrella trials were exploratory (n = 16/18), but the use of randomization was more common (n = 8/18). The median sample size of umbrella trials was 346 participants (IQR: 565–252 = 313), and the median study duration was 60.9 (IQR: 81.3–46.9 = 34.4) months. The median number of interventions investigated in umbrella trials was 5 (IQR: 6–4 = 2). The majority of platform trials were randomized (n = 15/16), and phase III investigation (n = 7/15; one did not report information on phase) was more common in platform trials with four of them using seamless II/III design. The median sample size was 892 (IQR: 1835–255 = 1580), and the median study duration was 58.9 (IQR: 101.3–36.9 = 64.4) months. Conclusions We anticipate that the number of master protocols will continue to increase at a rapid pace over the upcoming decades. More efforts to improve awareness and training are needed to apply these innovative trial design methods to fields outside of oncology.

Concurrent Surrogate Model Selection (COSMOS) Based on Predictive Estimation of Model Fidelity

2014 ◽  
Author(s):  
Souma Chowdhury ◽  
Ali Mehmani ◽  
Achille Messac
Keyword(s):  
Model Selection ◽  
Surrogate Model ◽  
Selection Process ◽  
Benchmark Problems ◽  
Optimal Model ◽  
Model Fidelity ◽  
Model Search ◽  
Binary Coding ◽  
Selection Framework ◽  
Optimal Kernel

One of the primary drawbacks plaguing wider acceptance of surrogate models is their low fidelity in general. This issue can be in a large part attributed to the lack of automated model selection techniques, particularly ones that do not make limiting assumptions regarding the choice of model types and kernel types. A novel model selection technique was recently developed to perform optimal model search concurrently at three levels: (i) optimal model type (e.g., RBF), (ii) optimal kernel type (e.g., multiquadric), and (iii) optimal values of hyper-parameters (e.g., shape parameter) that are conventionally kept constant. The error measures to be minimized in this optimal model selection process are determined by the Predictive Estimation of Model Fidelity (PEMF) method, which has been shown to be significantly more accurate than typical cross-validation-based error metrics. In this paper, we make the following important advancements to the PEMF-based model selection framework, now called the Concurrent Surrogate Model Selection or COSMOS framework: (i) The optimization formulation is modified through binary coding to allow surrogates with differing numbers of candidate kernels and kernels with differing numbers of hyper-parameters (which was previously not allowed). (ii) A robustness criterion, based on the variance of errors, is added to the existing criteria for model selection. (iii) A larger candidate pool of 16 surrogate-kernel combinations is considered for selection — possibly making COSMOS one of the most comprehensive surrogate model selection framework (in theory and implementation) currently available. The effectiveness of the COSMOS framework is demonstrated by successfully applying it to four benchmark problems (with 2–30 variables) and an airfoil design problem. The optimal model selection results illustrate how diverse models provide important tradeoffs for different problems.

Sign in / Sign up

Export Citation Format

Share Document