Dear all,

This is a reminder that the Reading Local Group meeting this autumn will be
part of RSS2000, held at Reading University.  Part of the agenda this
Wednesday, (September 13th),  has been organised by the local group
committee, and members are invited to attend this session, free of charge.
Details are given below.



> Kind regards
> 
> Reading Local Group Committee
> 
> 
> 
> 
> 
> Statistical Techniques for New Product Development
> Wednesday, September 13th
> Palmer Building, Room 101 (15.50 - 17.10)
> Chair: Peter Chapman (Zeneca Agrochemicals) 
> 
> 15.50-16.30
> "An overview of response surface methodology for new product development"
> Steven Gilmour (Queen Mary and Westfield College) 
> 
> 16.30-16.50
> "Examples of experimental design for new product development"
> Derek Pike (AB Statistics International & The University of Reading) 
> 
> 16.50-17.10
> "A practical application of mixtures experiments in chemical formulation
> development"
> Polly Martin (Zeneca Agrochemicals & The University of Reading) 
> 
> This will be followed by a wine reception and poster session in Room 104.
> 
> 
> Abstracts from these talks are supplied below, for your information:
> 
> An Overview of Response Surface Methodology for New Product Development
> Steven Gilmour
> School of Mathematical Sciences
> Queen Mary and Westfield College
> 
> Experimentation plays a crucial role in the development of new products,
> from basic research through to market testing.  The usefulness of any
> experiment depends on the information in the data which are collected from
> the experiment, which in turn depends on how the experiment was designed.
> It is therefore vital that the statistical principles of design are
> considered along with all other practical issues in planning experiments
> for new product development. 
> 
> One of the most important design ideas is to study several factors in the
> same experiment. If some of these factors are continuous, response surface
> methodology (RSM) is appropriate. RSM is a set of techniques for designing
> experiments, analysing the resulting data and interpreting fitted response
> surfaces. Important objectives typically include optimising one or more
> characteristics of the product and understanding how the factors affect
> those characteristics. RSM is used very widely by non-statisticans, but is
> not part of many statisticians' training. 
> 
> In this paper I will present a broad overview of RSM, describing the basic
> ideas, explaining their connection with other commonly used statistical
> and computational techniques, reviewing some of the useful recent
> developments in methodology and presenting some possible future
> directions. My presentation will be based on problems arising in the food,
> chemicals and pharmaceuticals industries, but RSM can also be useful in
> product development in manufacturing industry, agriculture and even
> clinical trials.
> 
> 
> 
> Experimental Design for New Product Development
> Derek J Pike
> AB Statistics International &
> Department of Applied Statistics, The University of Reading
> 
> It is critical for all industries today to be able to maintain or increase
> market share - either to remain ahead of the competition or to increase
> competitiveness with a current market leader.  One aspect of this activity
> is the continual drive to develop successful new products - and to do so
> in a way which is cost-effective.  
> 
> It is sad that so many industries are unaware both that the statistician
> has something to offer in this work, and that the tools to do the job have
> been there for years.
> 
> This paper will concentrate on new product development for the food and
> drink industry - an area which is different from some others in one
> important aspect.  New prototype products are invariably expensive to
> produce; but in this field they are relatively cheap to evaluate, and
> replicate data is never a problem to obtain.  This in itself raises
> additional design questions of interest, about how to set replication
> levels in relation to company objectives.
> 
> This paper re-evaluates the potential of sequential fractional factorial
> design as a tool for new product development, and presents a strategy in
> relation to a current practical problem.
> 
> 
> 
> 
> New Product Development in the Agrochemical Industry - A Case Study
> Polly Martin
> Department of Applied Statistics, The University of Reading
> 
> The development of a new agrochemical formulation has many stages; from
> the discovery and efficacy testing of an active ingredient, through
> formulation development, to regulatory testing and approval.
> 
> Formulation development involves experimentation to optimise physical
> properties of the formulation (for example viscosity, friability,
> dispersion time, particle size etc.) by changing the levels of some
> ingredients in the formulation.  In general by this stage of product
> development, the level of active ingredient is fixed, and the
> experimentation involves changing only the levels of three or four
> surfactants or other additives in the formulation.
> 
> It is common in this type of formulation development for there to be a
> make-up ingredient (often water or an organic solvent in a liquid
> formulation, or talc, clay or similar material in a solid formulation)
> that takes the total volume of the formulation to a set amount.
> Traditionally this make-up ingredient has been ignored in the context of
> the design and analysis of these experiments, although it can be included
> and the problem then thought of as a component proportion mixture problem.
> 
> In this paper I present an example of this type of formulation development
> and investigate the different options for tackling the problem.  A mixture
> approach is compared to both a slack-variable approach and an additive
> blending approach.
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%