Search

© 2015 Elsevier Ltd. In this paper, we analyze the origin-destination airport choice for freighter operations of combination and all-cargo carriers in Europe. First, we discuss the choice process of airlines qualitatively. Next, using a stated choice experiment, we show that the presence of forwarders at an airport is the primary factor in explaining airlines' choices, especially for airlines serving main airports. For airlines primarily serving regional airports, the possibility for night-time flights is most important. Finally, the presence of passenger operations at an airport is not a significant factor and the level of origin-destination demand is of limited importance. ispartof: Transportation Research E, Logistics and Transportation Review vol:87 pages:53-74 status: published

The effect of wheat bran on starch gelatinization temperature was investigated. Dynamic water vapour sorption and water retention capacity experiments showed that bran bound up to 3 times more water than starch. However, examining starch gelatinization in starch-bran-water mixtures with differential scanning calorimetry showed that the effect of substituting starch by bran differed from that of moving into a regime of limiting water. Modelling the effect of the mixture composition on starch gelatinization behavior indicated that the onset (To) and peak (Tp) gelatinization temperatures were positively impacted by the bran concentration in water. The conclusion temperature (Tc) was negatively affected by the water content. Fractionation experiments demonstrated that the increased To and Tp were mainly caused by the extractable wheat bran components, such as potassium and phosphorus, which decrease the plasticization capacity of the solvent. The mechanism behind our observations was explained with the side-chain liquid-crystalline polymeric model for starch.

When a multifactor experiment is carried out over a period of time, the responses may depend on a time trend. Unless the tests of the experiment are conducted in proper order, the time trend has a negative impact on the precision of the estimates of the main effects, the interaction effects, and the quadratic effects. A proper run order, called a trend-robust run order, minimizes the confounding between the effects’ contrast vectors and the time trend’s linear, quadratic, and cubic components. Finding a trend-robust run order is essentially a permutation problem. We develop a multistage approach based on integer programing to find a trend-robust run order for any given design. The multistage nature of our algorithm allows us to prioritize the trend robustness of the main-effect estimates. In the literature, most of the methods used are tailored to specific designs and are not applicable to an arbitrary design. Additionally, little attention has been paid to trend-robust run orders of response surface designs, such as central composite designs, Box–Behnken designs, and definitive screening designs. Our algorithm succeeds in identifying trend-robust run orders for arbitrary factorial designs and response surface designs with two up to six factors. ispartof: Journal Of Quality Technology vol:51 issue:1 pages:37-50 status: published

Data from choice experiments are analyzed using choice models. Proper modeling allows a good understanding of the consumer preferences, the identification of consumer segments and the optimization of products, and, eventually, offers opportunities for competitive advantage. However, in the analysis of choice data, products that are mixtures of ingredients have been largely overlooked. In this paper, we combine traditional choice models and traditional mixture models. We apply the resulting mixture choice models to data from a real-life experiment in which consumers made pair-wise comparisons involving seven cocktails. For the choice models, we first assume consumer homogeneity. Next, we allow for heterogeneity among individuals. Therefore, we discuss the multinomial logit model, the mixed logit model and the latent class model. For identifying segments, besides the latent class model, we explore various two-stage segmentation approaches in which forces, hierarchical Bayes estimates and Firth individual-level estimates are used as input for a cluster analysis. The results show that the mixed logit model and the latent class model describe the data better than the multinomial logit model. The optimal cocktails for the segments we identified using the various methods are not identical, but show many similarities. ispartof: FOOD QUALITY AND PREFERENCE vol:77 pages:135-146 status: published

Row–column designs involve two crossed blocking factors. We propose three combinatorial criteria and a criterion based on D-efficiency to identify optimal row–column arrangements of fractional factorial two-level designs. We search for such arrangements in complete catalogs of 16-run and 24-run orthogonal arrays of strength 2 with two-level treatment factors and two multi-level blocking factors. One of the combinatorial criteria is a good surrogate for the efficiency-based criterion. For searching similar catalogs of strength-3 orthogonal arrays, we adapt the criteria and conduct a complete search for optimal row–column arrangements in catalogs with 64 and 72 runs. © 2019, © 2019 American Society for Quality.

PCR enables sensitive and specific detection of infectious disease agents, but application in point-of-care diagnostic testing remains scarce. A compact tool that runs PCR assays in less than a few minutes and that relies on mass-producible, disposable reactors could revolutionize while-you-wait molecular testing. We here exploit well-established semiconductor manufacturing processes to produce silicon ultra-fast quantitative PCR (UF-qPCR) chips that can run PCR protocols with limited assay optimization. A total of 110 clinical samples were analyzed for the detection of group B streptococci using both a validated benchtop and an on-chip qPCR assay. For the on-chip assay, the total reaction time was reduced after optimization to less than 5 min. The standard curve, spanning a concentration range of 5 log units, yielded a PCR efficiency of 94%. The sensitivity obtained was 96% (96/100; CI: 90-98%) and the specificity 70% (7/10; CI: 40-90%). We show that if melting analyses would be integrated, the obtained sensitivity would drop slightly to 93% (CI: 86-96%), while the specificity would increase to 100% (CI: 72% - 100%). In comparison to the benchtop reference qPCR assay performed on a LightCycler©96, the on-chip assay demonstrated a highly significant qualitative (Spearman's rank correlation) and quantitative (linear regression) correlation. Using a mass-producible qPCR chip and limited assay optimization, we were able to develop a validated qPCR protocol that can be carried out in less than five minutes. The analytical performance of the microchip-based UF-qPCR system was shown to match that of a benchtop assay. This is the first report to provide UF-qPCR validation using clinical samples. We demonstrate that qPCR-based while-you-wait testing is feasible without jeopardizing assay performance. ispartof: TALANTA vol:192 pages:220-225 ispartof: location:Netherlands status: published

© 2015, Journal of Advertising Research. All rights reserved. The current study applied a “mixture-amount modeling” statistical approach—used most often in biology, agriculture, and food science—to measure the impact of advertising effort and allocation across different media. The authors of the current paper believe advertisers can use the mixture-amount model to detect optimal advertising-mix allocation changes as a function of their total advertising effort. The researchers demonstrated the use of the model by analyzing Belgian magazine and television data on 34 advertising campaigns for beauty-care brands. The goal is to help advertisers maximize desirable outcomes for campaign recognition and brand interest. ispartof: Journal of Advertising Research vol:55 issue:4 pages:443-457 status: published

© Copyright 2015 by ASQ. In some designed experiments, measurements of characteristics of the experimental units may be available prior to performing the runs. If the investigators believe that these measured characteristics may have some effect on the response of interest, then it seems natural to include these characteristics as factors in the experiment even though they are not under direct control. It may also be possible to apply multiple runs to a given experimental unit by subdividing it into multiple pieces, each having the same characteristics. A similar scenario involves using a person or an animal as an experimental unit multiple times but with different treatments. Here, the measured information about the subjects may not change over the experiment. In either of these cases, the fact that several runs employ the same experimental unit means that the responses for those runs are correlated. This correlation, in addition to the natural variability of the measured characteristics over the sample of available experimental units, requires new methodology for creating optimal designs. Specifically, the methodology must choose a subset of the experimental units and determine the number of treatments applied to each experimental unit in addition to choosing the level combinations of the controllable factors for each run. In this article, we provide two methods for generating optimal designs in the presence of additional information about the experimental units. The first method fixes the number of runs performed on each experimental unit. The second method allows for varying numbers of runs applied to each experimental unit, subject to a constraint on the total number of runs. We discuss several illustrative examples using each method as well as a real experiment using previously fabricated batches of polypropylene as experimental units in a study on the effects of a subsequent plasma treatment. ispartof: Journal of Quality Technology vol:47 issue:4 pages:301-317 status: published

BACKGROUND: The selection of a future breeding dog is a complicated task, in which disease characteristics and different traits have to be combined and weighed against one another. Truncation selection, that is the exclusion of affected animals, may be very inefficient when selecting on a large number of traits, and may result in a reduction of the genetic diversity in a population or breed. Selection could be facilitated by the use of a selection index that combines multiple traits or breeding values into one score. This however requires a consideration of their relative value according to their economic weight, which is difficult to express in monetary units for health traits. The use of a choice experiment to derive non-market values might be a solution to this problem. This is a pilot study to assess the potential use of choice experiments to ascertain the public preference and relative importance attached to health- and conformation traits in the selection of a Cavalier King Charles spaniel. The focus was on two prevalent disorders, mitral valve disease and syringomyelia, and on several important conformation traits such as muzzle length and eye shape. Based on available prior information, a Bayesian D-optimal design approach was used to develop a choice experiment and the resulting choice sets. RESULTS: Every participant (breeder or owner) in the choice experiment was presented with a total of 17 choice sets, in which at most four traits could vary to reduce the cognitive burden. A total of 114 respondents participated in the choice experiment and results showed that respondents (breeders/owners) current attitudes were directed towards health (syringomyelia and mitral valve disease), followed by eye shape and level of inbreeding. CONCLUSIONS: This approach identifies the value breeders and owners attach to certain traits in the breeding objective. The resulting relative weights, represented as the logworths obtained from the choice experiment, could be an alternative to economic weights. They could be implemented as a weight when breeding values are available, but more study on this topic will be necessary. A challenge in this approach is to scale up the experiment with additional traits. Moreover, for other traits, the genetic parameters and correlations should be known first, in order to include them in the health selection index as well. ispartof: Canine Genetics and Epidemiology vol:5 pages:1-7 ispartof: location:England status: published

© 2018 Elsevier Ltd Two-level orthogonal designs play an important role in industrial screening experiments, in which the primary goal is to identify the treatment factors with the largest main effects on the output of a process or the performance of a product. Often, the experimental tests suggested by the orthogonal designs cannot be performed on a single day or using a single batch of raw material. This causes day-to-day or batch-to-batch variation in the experimental results, and necessitates the use of orthogonal blocking patterns. These blocking patterns ensure that the factors’ main effects can be estimated independently from the day-to-day or batch-to-batch variation. Finding an optimal orthogonal blocking pattern for an orthogonal design is a major challenge. Recently, mixed integer linear programming has been used for that purpose. While this approach is promising, computational experiments have indicated it is quite slow. We show how to speed up the mixed integer linear programming approach by adding symmetry breaking constraints to the formulation, and study the usefulness of an asymmetric representatives formulation. In other words, we introduce state-of-the-art symmetry breaking approaches in optimal experimental design. We perform extensive computational experiments to test which combinations of symmetry breaking constraints work best. Throughout, we consider two kinds of symmetry: symmetry due to the fact that the blocks can be relabeled without affecting the quality of the blocking pattern, and symmetry due to replicated test combinations. ispartof: COMPUTERS & OPERATIONS RESEARCH vol:97 pages:96-110 status: published

The best demonstration of the ruggedness of a method is monitoring its performance on an ongoing basis as part of the analytical quality control applied in the laboratory. However, an initial demonstration of the ruggedness is often performed as one aspect of the method validation, to give confidence that the method should perform well under normal variations in conditions in routine application. This initial ruggedness testing is typically performed using either multiple replicate analyses or application of design of experiments (DoEs) which minimizes the number of analyses, time and effort required to detect influences on the measurement results. Two DoEs were applied for ruggedness testing for a modified QuEChERS multiresidue method for the detection of pesticide residues in potato by GC-MS/MS. Seven experimental factors were studied using an eight-run Plackett-Burman design replicated three times and an augmented definitive screening design with 34 runs. The relative effectiveness of the two approaches is discussed, in terms of their statistical significance, their cost-effectiveness and the richness of information they provide on the effects of the parameters investigated and the actual robustness of the method being tested.

© 2014 Elsevier Ltd. Due to the globalization and the fragmentation of industrial production processes, the logistics sector, organizing the linkages between different production plants and the market, is growing fast. This results in an increasing demand for suitable new business locations. Previous research has indicated that accessibility is a key factor in the location decision making process. Though the literature on this subject is extensive, little research has been done to quantify the impact of the different dimensions of accessibility on the location decision process of logistics companies. This paper aims to fill this void in the literature by means of a revealed preference study (using a Geographic Information System (GIS) analysis) and a stated preference study (using a designed discrete choice experiment) in Flanders (Belgium). The results of the revealed preference study served as input to the design of the choice situations in the stated preference study. In the stated preference study, the respondents were confronted with a series of choice situations described by means of accessibility variables as well as land rent information. An analysis of the resulting data by means of discrete choice modeling revealed that land rent is the most important factor in the location choice of logistics companies in Flanders. Access to a port is the second most important factor, followed by access to a motorway, the location in a business park and an inland navigation terminal, which are all about equally important. Access to a rail terminal plays no significant role in the location choice of logistics companies in Flanders. ispartof: Journal of Transport Geography vol:42 pages:110-121 status: published

Food chemistry 264, 367 - 376 (2018). doi:10.1016/j.foodchem.2018.04.125, Chelation of iron and zinc in wheat as phytates lowers their bio-accessibility. Steeping and germination (15 °C, 120 h) lowered phytate content from 0.96% to only 0.81% of initial dry matter. A multifactorial experiment in which (steeped/germinated) wheat was subjected to different time (2–24 h), temperature (20–80 °C) and pH (2.0–8.0) conditions showed that hydrothermal processing of germinated (15 °C, 120 h) wheat at 50 °C and pH 3.8 for 24 h reduced phytate content by 95%. X-ray absorption near-edge structure imaging showed that it indeed abolished chelation of iron to phytate. It also proved that iron was oxidized during steeping, germination and hydrothermal processing. It was further shown that zinc and iron bio-accessibility were respectively 3 and 5% in wheat and 27 and 37% in hydrothermally processed wheat. Thus, hydrothermal processing of (germinated) wheat paves the way for increasing elemental bio-accessibility in whole grain-based products., Published by Elsevier, New York, NY [u.a.]

In many discrete choice experiments set up for product innovation, the number of attributes is large, which results in a substantial cognitive burden for the respondents. To reduce the cognitive burden in such cases, Green suggested in the early '70s the use of partial profiles that vary only the levels of a subset of the attributes. In this paper, we present two new methods for constructing Bayesian D-optimal partial profile designs for estimating main-effects models. They involve alternative generalizations of Green's approach that makes use of balanced incomplete block designs and take into account the fact that attributes may have differing numbers of levels. We refer to our methods as variance balance I and II because they vary an attribute with a larger number of levels more often than an attribute with fewer levels to stabilize the variances of the individual part-worth estimates. The two variance balance methods differ in the way attributes with differing numbers of levels are weighted. Both methods provide statistically more efficient partial profile designs for differing numbers of attribute levels than another generalization of Green's approach that does not weight the attributes. This method is called attribute balance. We show results from an actual experiment in software development demonstrating the usefulness of our methods. Copyright © 2014 John Wiley & Sons, Ltd.

This paper revisits the problem of finding continuous V-optimal mixture designs for the qth degree model. For this problem, Liu and Neudecker [10] present general analytical expressions for the weights of the points of the simplex-centroid design and claim that the resulting designs are V-optimal. For a problem involving three mixture ingredients, Liu and Neudecker's results are in conflict with those of Laake [6]. We find that Laake's design for three ingredients is superior to that of Liu and Neudecker, and we establish the V-optimality of Laake's design and report V-optimal weights for the points of the simplex-centroid design for cases involvingmore than three ingredients. All newly proposed designs are superior to those of Liu and Neudecker.

© 2018 Microalgal lipids have a huge potential in several applications such as biofuels, food and feed. However, enzymatic lipolytic reactions during wet storage cause the formation of substantial amounts of free fatty acids (FFA), which give several problems for these applications. In this study, a full factorial (32) design with varying time and temperature was used to optimize a heat treatment, that inactivates lipolytic enzymes in the wet paste of T-Isochrysis lutea. Therefore, the FFA and lipid content were evaluated at three time points (0, 3 and 7 days) during wet storage at 4 °C. A treatment of at least 8 min at 80 °C was found to inactivate the lipolytic enzymes. A second experiment elucidated that this optimal treatment had no negative impact on fatty acid composition, peroxide value, phenolic content or carotenoid content. A short heat treatment, thus, enables wet storage of microalgal paste to several days at 4 °C. ispartof: ALGAL RESEARCH-BIOMASS BIOFUELS AND BIOPRODUCTS vol:38 status: published

Econometric Institute, Erasmus School ofEconomics, Erasmus University Rotterdam, Postbus 1738, 3000 DR Rotterdam, the Netherlands*Corresponding author. Email: tim.engels@uantwerpen.beAssessments of research groups by expert panels are commonplace. A conceptual distinctionbetween nationwide evaluations, approaches fitting within a national framework, and approachesfulfilling a national, regional, or local regulatory requirement can be made. Few papers exploit thewealth of the data made available to panels: most analyses and discussions focus on the relationbetween assessments scores and bibliometric indicators. However, expert panels are expected totake other information such as funding and PhDs into account as well. Using a data set on 52science research groups that have been evaluated in view of an on-going series of researchassessments at the University of Antwerp (Belgium), we build cumulative logistic regressionmodels that explain the assessments of research group quality and productivity. In addition tothe discipline, the predictors in the models explaining quality and productivity are group size,h-index of the group leader, and efficiency in terms of publishing in top journals. Strikingly, thesame predictors apply in the model for quality and in the model for productivity.Keywords: peer review; group size; research assessment exercise; research quality; researchproductivity.

This paper presents a general method for finding optimal blocking arrangements of pure-level and mixed-level orthogonal designs of resolution 3. The method requires enumeration of all nonisomorphic 'full designs' that include the treatment factors as well as the blocking factor. For all these designs, generalized word counts expressing the aliasing between main effects and two-factor interactions as well as the aliasing among two-factor interactions are calculated. The same is done for all projections into treatment designs obtained by dropping any possible blocking factor. The generalized word counts of the full design and treatment designs then allow the selection of blocking arrangements that are optimal with respect to five criteria appropriate for blocking resolution-3 orthogonal designs. We provide optimal blocking arrangements for orthogonal pure-level and mixed-level designs of 12, 16, 18, 20, and 27 runs.

Problem: In the field of genomics, proteomics and disease diagnostics, it is of great interest to capture individual targets (e.g., cells, bacteria, exosomes) on magnetic beads and to seed them in a microwell array. Optical manipulation of the magnetic beads in the array then allows selecting the targets of interest and moving them to another location for post analysis. However, this technology is hampered by the low fraction of beads that show Brownian motion within their microwell and thus can be optically trapped and manipulated. This is due to the fact that magnetic beads seeded in a microwell array are subject to attraction forces that cause them to adhere to the microwell surfaces. The challenge is to identify buffer conditions that prevent this from happening. Approach: We investigated five factors that were believed to influence the fraction of magnetic beads that show Brownian motion within their microwell. To this end, we used an experimental approach to maximize the fraction of beads showing Brownian motion. First, an initial D-optimal five-factor completely randomized design was created, including specific control treatments, for the estimation of a response surface model including a two-level and a six-level categorical factor (bead type and type of surfactant). Subsequently, two I-optimal follow-up experiments were conducted to examine the effectiveness of higher levels for one of the quantitative factors under investigation. Results: The initial D-optimal five-factor design indicated one insignificant quantitative factor. The remaining four factors all had a significant main effect. In addition, there were three significant interaction effects and one significant quadratic effect. The analysis of the data from the initial experimental design also revealed a strongly positive effect for two of the six surfactants. In the follow-up experiment, these two surfactants turned out to produce even larger predicted responses at higher concentrations. ispartof: Journal of Quality Technology vol:49 issue:4 pages:402-417 status: published

© 2017 Taylor & Francis. Many industrial experiments involve some factors that are hard to change. In this situation, experimenters often choose to perform an experiment with restricted randomization, such as a split-plot or a strip-plot experiment. In this article, we discuss the analysis of an experiment concerning the adhesion between steel tire cords and rubber. Besides an ordinal response, the experiment also involves one hard-to-change factor. Therefore, the experimenters performed a split-plot experiment. An additional complication of the experiment is that there is also a blocking factor. A proper analysis of the experiment requires the inclusion of random effects in the model to account for its split-plot nature and its blocked nature. The need for random effects and the ordinal response necessitate the use of a mixed cumulative logit model. ispartof: Quality Engineering vol:29 issue:4 pages:553-562 status: published

© 2017 American Statistical Association. The potential of two-level orthogonal designs to fit models with main effects and two-factor interaction effects is commonly assessed through the correlation between contrast vectors involving these effects. We study the complete catalog of nonisomorphic orthogonal two-level 24-run designs involving 3–23 factors and we identify the best few designs in terms of these correlations. By modifying an existing enumeration algorithm, we identify the best few 28-run designs involving 3–14 factors and the best few 36-run designs in 3–18 factors as well. Based on a complete catalog of 7570 designs with 28 runs and 27 factors, we also seek good 28-run designs with more than 14 factors. Finally, starting from a unique 31-factor design in 32 runs that minimizes the maximum correlation among the contrast vectors for main effects and two-factor interactions, we obtain 32-run designs that have low values for this correlation. To demonstrate the added value of our work, we provide a detailed comparison of our designs to the alternatives available in the literature. Supplementary materials for this article are available online. ispartof: Journal of the American Statistical Association vol:112 issue:519 pages:1354-1369 status: published

© 2017 Elsevier B.V. A crucial stage in the development of medical drugs is to study which additives, usually called excipients, impact the active ingredient stability. This type of study is generally named an excipient compatibility study and requires a mixture experiment. Subsequently, the effect of the storage conditions, more specifically the relative humidity and temperature, on the stability is investigated. This so-called accelerated life test involves a factorial type of experiment. It has become, however, customary to include the storage conditions in the compatibility study. This provides valuable information concerning potential interactions between excipient combinations and storage conditions. Experiments that combine a mixture experiment with a factorial experiment are generally named mixture-process variable experiments. A limited number of designs for mixture-process variable experiments are available in the literature. One problem is that the proposed designs offer little flexibility. Another is that the required number of runs becomes prohibitively large for large numbers of mixture components. In this paper, we examine flexible, optimal designs for realistic mixture-process variable experiments. Our motivation is to provide guidance to pharmaceutical formulation scientists concerning state-of-the art models and designs for excipient compatibility studies. Using several proof-of-concept examples, we demonstrate that I-optimal designs offer both flexibility and small variances of prediction. We also discuss a real-life example, which could be used as a blueprint for future studies. Because many excipient compatibility studies are not completely randomized, we pay special attention to their logistics and to the resulting randomization restrictions, which lead to split-plot and strip-plot experiments. ispartof: Chemometrics and Intelligent Laboratory Systems vol:171 pages:125-139 status: published

© 2015 Advertising Association. We investigate the moderating role of national openness to foreign markets on consumer responses (attitude to ad, brand and purchase intention) to different degrees of advertising adaptation in Belgium, Iran and India. We operationalize three levels of advertising adaptation (standardization, glocalization and adaptation) by manipulating the model (international versus local celebrity) and advertising copy (international versus local advertising copy). The results show that societies with a low openness to foreign markets respond more positively to fully adapted ads than to glocalized and standardized ads. The differences in responses decrease with national openness. peerreview_statement: The publishing and review policy for this title is described in its Aims & Scope. aims_and_scope_url: http://www.tandfonline.com/action/journalInformation?show=aimsScope&journalCode=rina20 ispartof: International Journal of Advertising vol:36 issue:2 pages:293-313 status: published

Rheology of cake batter is considered an important parameter determining cake quality. It is thus frequently used to investigate structural changes in the process of cake making. However, little is known about the relation between batter rheological properties and cake’s final properties, concerning volume and texture. In order to explore this relation an extensive experiment covering broad recipe variability was performed using an advanced experimental design approach and subsequent statistical analysis. The mixture of standard with pregelatinized wheat flour had significant main and interaction effects on most responses characterizing cake batter. Leavening acids and mixing time affected only a few of them. The simultaneous influence of ingredients and mixing time was clearly described. Measuring the batter rheological properties immediately after mixing gave sufficient information to significantly estimate in advance the volume of the cakes together with their cohesiveness. This methodology could improve the design of new bakery products with desirable volume and texture. ispartof: Journal of Cereal Science vol:77 pages:219-227 status: published

Strip-plot designs are commonly used in situations in which the production process consists of two process stages involving hard-to-change factors and in which it is possible to apply the second stage to semifinished products from the first stage. This article focuses on three-stage processes. As opposed to the three-stage strip-plot designs in the literature, the third stage does not involve hard-to-change factors; instead, it involves easy-to-change factors that are reset independently for each run. For this scenario, the split-split-plot design is a well-known alternative design option. However, the authors prefer the more statistically efficient strip-plot designs and, therefore, they construct D-optimal strip-plot designs for three-stage processes with no randomization restriction in the third stage. The coordinate-exchange algorithm used to construct the designs can handle any type of factor and any number of factor levels, runs, rows, and columns.

In many industrial experiments, complete randomization of the runs is impossible as, often, they involve factors whose levels are hard or costly to change. In such cases, the split-plot design is a cost-efficient alternative that reduces the number of independent settings of the hard-to-change factors. In general, the use of generalized least squares is required for model estimation based on data from split-plot designs. However, the ordinary least squares estimator is equivalent to the generalized least squares estimator for some split-plot designs, including some second-order split-plot response surface designs. These designs are called equivalent-estimation designs. An important consequence of the equivalence is that basic experimental design software can be used for model estimation. This article introduces two new families of equivalent-estimation split-plot designs, one based on subset designs and another based on supplementary difference set designs. The resulting designs complement existing catalogs of equivalent-estimation designs and allow for a more flexible choice of the number of hard-to-change factors, the number of easy-to-change factors, the number and size of whole plots, and the total sample size. It is shown that many of the newly proposed designs possess good predictive properties when compared to D-optimal split-plot designs.

A novel meta-heuristic approach for minimizing nonlinear constrained problems is proposed, which offers tolerance information during the search for the global optimum. The method is based on the concept of design and analysis of computer experiments combined with a novel two phase design augmentation (DACEDA), which models the entire merit space using a Gaussian process, with iteratively increased resolution around the optimum. The algorithm is introduced through a series of cases studies with increasing complexity for optimizing uniformity of a short-wave infrared (SWIR) hyperspectral imaging (HSI) illumination system (IS). The method is first demonstrated for a two-dimensional problem consisting of the positioning of analytical isotropic point sources. The method is further applied to two-dimensional (2D) and five-dimensional (5D) SWIR HSI IS versions using close- and far-field measured source models applied within the non-sequential ray-tracing software FRED, including inherent stochastic noise. The proposed method is compared to other heuristic approaches such as simplex and simulated annealing (SA). It is shown that DACEDA converges towards a minimum with 1 % improvement compared to simplex and SA, and more importantly requiring only half the number of simulations. Finally, a concurrent tolerance analysis is done within DACEDA for to the five-dimensional case such that further simulations are not required. ispartof: Optics Express vol:24 issue:26 pages:29380-29405 ispartof: location:United States status: published

© 2016 John Wiley & Sons Ltd. This study was set up to identify the long-term effect of biochar on soil C sequestration of recent carbon inputs. Arable fields (n = 5) were found in Belgium with charcoal-enriched black spots (>50 m2; n = 14) dating >150 years ago from historical charcoal production mound kilns. Topsoils from these 'black spots' had a higher organic C concentration [3.6 ± 0.9% organic carbon (OC)] than adjacent soils outside these black spots (2.1 ± 0.2% OC). The soils had been cropped with maize for at least 12 years which provided a continuous input of C with a C isotope signature (δ13C) -13.1, distinct from the δ13C of soil organic carbon (-27.4 ‰) and charcoal (-25.7 ‰) collected in the surrounding area. The isotope signatures in the soil revealed that maize-derived C concentration was significantly higher in charcoal-amended samples ('black spots') than in adjacent unamended ones (0.44% vs. 0.31%; P = 0.02). Topsoils were subsequently collected as a gradient across two 'black spots' along with corresponding adjacent soils outside these black spots and soil respiration, and physical soil fractionation was conducted. Total soil respiration (130 days) was unaffected by charcoal, but the maize-derived C respiration per unit maize-derived OC in soil significantly decreased about half (P < 0.02) with increasing charcoal-derived C in soil. Maize-derived C was proportionally present more in protected soil aggregates in the presence of charcoal. The lower specific mineralization and increased C sequestration of recent C with charcoal are attributed to a combination of physical protection, C saturation of microbial communities and, potentially, slightly higher annual primary production. Overall, this study provides evidence of the capacity of biochar to enhance C sequestration in soils through reduced C turnover on the long term. ispartof: Global Change Biology Bioenergy vol:8 issue:2 pages:371-381 status: published

To calibrate spectral data, one typically starts with preprocessing the spectra and then applies a multivariate calibration method such as principal component regression or partial least squares regression. In the model selection step, the optimal number of latent variables is determined in order to minimize the prediction error. To protect the analysis against the harmful influence of possible outliers in the data, robust calibration methods have been developed. In this paper, we focus on the preprocessing and the model selection step. We propose several robust preprocessing methods as well as robust measures of the root mean squared error of prediction (RMSEP). To select the optimal preprocessing method, we summarize the results for the different RMSEP values by means of a desirability index, which is a concept from industrial quality control. These robust RMSEP values are also used to select the optimal number of latent variables. We illustrate our newly developed techniques through the analysis of a real data set containing near-infrared measurements of samples of animal feed.

Response surface experiments often involve only quantitative factors, and the response is fit using a full quadratic model in these factors. The term response surface implies that interest in these studies is more on prediction than parameter estimation because the points on the fitted surface are predicted responses. When computing optimal designs for response surface experiments, it therefore makes sense to focus attention on the predictive capability of the designs. However, the most popular criterion for creating optimal experimental designs is the D-optimality criterion, which aims to minimize the variance of the factor e↵ect estimates in an omnibus sense. Because I-optimal designs minimize the average variance of prediction over the region of experimentation, their focus is clearly on prediction. Therefore, the Ioptimality criterion seems to be a more appropriate one than the D-optimality criterion for generating response surface designs. Here we introduce I-optimal design of split-plot response surface experiments. We show through several examples that I-optimal split-plot designs provide substantial benefits in terms of improved prediction compared with D-optimal split-plot designs, while also performing very well in terms of the precision of the factor e↵ect estimates.

In a discrete choice experiment, each respondent chooses the best product or service sequentially from many groups or choice sets of alternative goods. The alternatives are described by levels of a set of prede ned attributes and are also referred to as pro les. Respondents often nd it dicult to trade o prospective goods when every attribute of the oering changes in each comparison. Especially in studies involving many attributes, respondents get overloaded by the complexity of the choice task. To overcome respon- dent fatigue, it is better to simplify the choice tasks by holding the levels of some of the attributes constant in every choice set. The resulting designs are called partial pro le designs. In this paper, we construct D-optimal par- tial pro le designs for estimating main-eects models. We use a Bayesian design algorithm that integrates the D-optimality criterion over a prior dis- tribution of likely parameter values. To determine the constant attributes in each choice set, we generalize the approach that makes use of balanced incomplete block designs. Our algorithm is very exible because it produces partial pro le designs of any choice set size and allows for attributes with any number of levels and any number of constant attributes. We provide an illustration in which we make recommendations that balance the loss of statistical information and the burden imposed on the respondents. Keywords: discrete choice experiments; Bayesian D-optimal design; partial pro les; lexicographic choice behavior; attribute balance; coordinate-exchange algorithm

Designed experiments allow researchers to explore and understand the relation between sets of experimental factors and one or more responses. However, there are many cases where the experimenter cannot determine the factor levels freely. In this article, we study a baking experiment where the interest is in understanding the relationship between flour characteristics and bread quality. The experimental factors are six mixture components and one observable process variable, so that the baking experiment is a mixture-process variable (MPV) experiment. The starting point for the study is an available stock of 30 flour samples for which the mixture component proportions and the level of the process variable have been determined using an elaborate chemical analysis. The challenge for the experimenter is to determine which flour samples, or which combinations of flour samples, should be used in the actual experiment. Since the design region is highly constrained and the usable factor levels are dictated by the set of 30 initial flour samples, classical MPV designs are not applicable and an optimal design approach is required. We compare two different design strategies, one based on a point-exchange algorithm and another based on a coordinate-exchange algorithm. The coordinate-exchange algorithm leads to the most efficient design since it is not bounded by a given candidate set of points. However, it requires the calculation of the convex hull of the set of 30 flour samples along with the related constraints. Because of the high dimensionality of the problem, determining the convex hull and using it to create the constraints for the experimental design is a tedious process. In contrast, the point-exchange algorithm starts from a candidate set of possible factor level combinations, a limitation that lowers the efficiency of the design but results in a design that is very easy to implement.

In a rank-order choice-based conjoint experiment, the respondent is asked to rank a number of alternatives of a number of choice sets. In this paper, we study the efficiency of those experiments and propose a D-optimality criterion for rank-order experiments to find designs yielding the most precise parameter estimators. For that purpose, an expression of the Fisher information matrix for the rank-ordered conditional logit model is derived which clearly shows how much additional information is provided by each extra ranking step. A simulation study shows that, besides the Bayesian D-optimal ranking design, the Bayesian D-optimal choice design is also an appropriate design for this type of experiments. Finally, it is shown that considerable improvements in estimation and prediction accuracy are obtained by including extra ranking steps in an experiment.

This paper presents a systematic approach to dealing with within-pair order effects in paired comparison experiments, which have historically received substantial attention in many application areas, including sensometrics, psychology, transportation, marketing and medicine. It describes how optimal designs can be constructed for paired comparison experiments involving several attributes. It extends earlier work on design of experiments in the presence of within-pair order effects for a single qualitative attribute and fills a void in the recent design literature on choice experiments with multiple attributes.

Summary Split-plot and other multistratum structures are widely used in factorial and response surface experiments. Residual maximum likelihood (REML) and generalized least squares (GLS) estimation is seen as the state of the art method of data analysis for non-orthogonal designs. We analyse data from an experiment that was run to study the effects of five process factors on the drying rate for freeze-dried coffee and find that the main plot variance component is estimated to be 0. We show that this is a typical property of REML–GLS estimation in non-orthogonal split-plot designs with few main plots which is highly undesirable and can give misleading conclusions. Instead, we recommend a Bayesian analysis, using an informative prior distribution for the main plot variance component and implement this by using Markov chain Monte Carlo sampling. Paradoxically, the Bayesian analysis is less dependent on prior assumptions than the REML–GLS analysis. Bayesian analyses of the coffee freeze-drying data give more realistic conclusions than REML–GLS analysis, providing support for our recommendation.

Recently, Kessels et al. (2006) developed a way to produce Bayesian G- and V-optimal designs for the multinomial logitmodel. These designs allow for precise response predictions which is the goal of conjoint choice experiments. The authors showed that the G- and V- optimality criteria outperform the D- and A-optimality criteria for prediction. However, their G- and V-optimal design algorithm is computationally intensive, which is a barrier to its use in practice. In this paper, we present an efficient algorithm for calculating Bayesian optimal designs by means of the different criteria. Particularly, the speed of computation for the V-optimality criterion has improved dramatically.The new algorithm makes it possible to use Bayesian D-, A-, G- and V-optimal designs that are tailored to individual respondents in computerized conjoint choice studies.

markdownabstractMany products and services can be described as mixtures of ingredients whose proportions sum to one. Specialized models have been developed for linking the mixture proportions to outcome variables, such as preference, quality and liking. In many scenarios, only the mixture proportions matter for the outcome variable. In such cases, mixture models suffice. In other scenarios, the total amount of the mixture matters as well. In these cases, one needs mixture- amount models. As an example, consider advertisers who have to decide on the advertising media mix (e.g. 30% of the expenditures on TV advertising, 10% on radio and 60% on online advertising) as well as on the total budget of the entire campaign. To model mixture-amount data, the current strategy is to express the response in terms of the mixture proportions and specify mixture parameters as parametric functions of the amount. However, specifying the functional form for these parameters may not be straightforward, and using a flexible functional form usually comes at the cost of a large number of parameters. In this paper, we present a new modeling approach which is flexible but parsimonious in the number of parameters. The model is based on so-called Gaussian processes and avoids the necessity to a-priori specify the shape of the dependence of the mixture parameters on the amount. We show that our model encompasses two commonly used model specifications as extreme cases. Finally, we demonstrate the model’s added value when compared to standard models for mixture-amount data. We consider two applications. The first one deals with the reaction of mice to mixtures of hormones applied in different amounts. The second one concerns the recognition of advertising campaigns. The mixture here is the particular media mix (TV and magazine advertising) used for a campaign. As the total amount variable, we consider the total advertising campaign exposure.