Data Analysis and Mathematical Modeling
HaloteC strongly focuses on research and development of mathematical models and data analysis tools.
The amount of data that is generated by modern sensors or any data acquisition system, has increased dramatically during the past decades. Although in many cases data is stored on hard drives, the step
required to turn it into useful information is often not, or only partially taken.
Optimization of the data storage process (databases) and the availability of validated models contributes to:
- Increased understanding of complex process dynamics, which may provide a cost effective means
for process optimization, faster product development and saving (labor) cost - Sharing knowledge and expertise with customers and partners, e.g. through dedicated web interfaces
- Revealing hidden structures and information in large data sets, which may lead to more accurate
extrapolations and predictions
Numerical Toolbox
During the past decade, HaloteC has developed a general toolbox with numerical algorithms that can be used to solve mathematical models of complex (dynamic) systems, such as biotechnological processes (cell cultures) or specific sensor outputs. The toolbox contains a/o routines for solving large sets of (non-linear) algebraic
and/or differential equations and routines for optimization and statistical analysis.
During the past decade, HaloteC has developed a general toolbox with numerical algorithms that can be used to solve mathematical models of complex (dynamic) systems, such as biotechnological processes (cell cultures) or specific sensor outputs. The toolbox contains a/o routines for solving large sets of (non-linear) algebraic
and/or differential equations and routines for optimization and statistical analysis.
Evolutionary Algorithms
An intriguing development is the use of evolutionary algorithms for solving all kinds
of real world problems. The basic thought behind this approach is to generate a set
of random models, the initial population and use specific rules to create offspring.
If these child models perform better, they are kept in the population. The process continues until some criterion is satisfied. HaloteC focuses on developing evolutionary algorithms that continually run to find models that best describes a given
(large) data set.
An intriguing development is the use of evolutionary algorithms for solving all kinds
of real world problems. The basic thought behind this approach is to generate a set
of random models, the initial population and use specific rules to create offspring.
If these child models perform better, they are kept in the population. The process continues until some criterion is satisfied. HaloteC focuses on developing evolutionary algorithms that continually run to find models that best describes a given
(large) data set.
Calculation Cluster
Large or complex mathematical models may require a lot of computational power.
For CPU intensive calculations HaloteC has developed a very modular, Linux based calculation cluster. The cluster can be extended, simply adding more calculation nodes (CPU's). Complex calculation jobs, e.g. from web-calculator interfaces, can be
processed in parallel. Subsequently, results may be reported with auto-generated
pdf documents that are automatically send back to the user via e-mail.
Large or complex mathematical models may require a lot of computational power.
For CPU intensive calculations HaloteC has developed a very modular, Linux based calculation cluster. The cluster can be extended, simply adding more calculation nodes (CPU's). Complex calculation jobs, e.g. from web-calculator interfaces, can be
processed in parallel. Subsequently, results may be reported with auto-generated
pdf documents that are automatically send back to the user via e-mail.