Share full text access. Please review our Terms and Conditions of Use and check box below to share full-text version of article.
Abstract Genetic algorithms emulate biologic evolutionary concepts to solve search and optimization problems. Citing Literature.
You are here
Volume 13 , Issue 1 January Pages Related Information. Close Figure Viewer. Browse All Figures Return to Figure. Previous Figure Next Figure. Email or Customer ID. Forgot password? In medicine, the application of GAs is not limited to image processing, but includes very complex use in radiology, oncology, cardiology, endocrinology, gynecology, pediatrics, surgery, pulmonology, infectology, radiotherapy, rehabilitation medicine, orthopedics, neurology and pharmacology — hence in almost every branch of medicine 1.
Actually, at least one chapter could be written about the application of GA in each of these areas. GAs are machine-learning and optimization techniques that are used to solve complex search problems.
Genetic Algorithms: Concepts and Designs
In medical science and engineering, GAs are used to search for huge numbers of possible combinations of parameters in order to find the combination that best suits the realistic or desired endpoint. The basic flow of GA consists of five parts Figure 1. Through the initialization part, a random virtual population of a certain size is created.
Population size may vary from several to thousands of individuals. In the next step, the fitness of each individual is determined. Fitness is defined as the proportion in which an individual satisfies the property required by the algorithm. In the third step, this population is improved by keeping only those individuals that best match the required properties. Hence, this step is called a selection. The selection is followed by crossover, in which new individuals are created with the combinations of the existing, ie, survived individuals.
In this step, the assumption is to get the individuals that better fit the desired property. The crossover is followed by a mutation step, in which individuals undergo small random changes. After the mutation, the whole sequence is repeated, starting from the evaluation step until the population with the desired properties is obtained.
GAs usage in medicine can be explained using a very simplified case of therapy optimization. Let us suppose that for the treatment of a patient seven different therapies are available that can be combined and that do not interact with each other. Each therapy causes measurable damage to the organism. The chance of success of each therapy and the harm to the organism are shown in Table 1. The R script that can solve this problem is shown in Figure 2. As the mutation is random, each calculation provides a different output sequence.
The described example is a simplified presentation of application of GA for therapy optimization. However, sophisticated GA-based mathematical tools are already widely applied in medicine. Only in the past few years hundreds of scientific and professional articles have been published using some kind of GA.
For example, a neural network based on GA GANN — Genetic Algorithm Neural Network has in the late nineties been successfully used to predict the outcomes of non-small cell lung cancer surgical treatments 2. GANN systems constructed in this manner can predict far better than the methods based on the usual logistic regression. Thus, in , an article describing the use of GA to detect a hypoglycemic condition based on an EEG signal was published 3. GAs can be used in conjunction with other techniques, such as recursive local floating enhancement technique LFE 4.
Two key steps in GA, combination crossover and mutation, ensure that the entire population of each generation will go to the ideal, ie, needed stage. On the other hand, GA cannot improve the individual to its local optimum. In this way, by using such a hybrid GA, a greater number of combinations can be processed simultaneously.
Genetic Algorithms: Concepts and Designs
An example of the use of such a hybrid GA is the detection of seven protein markers to determine the risk of major adverse cardiac events 5. Another technique is supporting vectoring machines SVM algorithm, ie, supervised learning model with related learning algorithms that analyze the data used for regression analysis and classification. GA in conjunction with SVM was used to predict the cardiovascular fetal state 6. Upload PDF. Follow this author. New articles by this author.
Genetics without genes: application of genetic algorithms in medicine
New citations to this author. New articles related to this author's research. Email address for updates.
- Genetic Algorithm - MATLAB & Simulink.
- Citations per year!
- Genetic Algorithms: Concepts and Designs.
My profile My library Metrics Alerts. Sign in. Get my own profile Cited by View all All Since Citations h-index 43 29 iindex Verified email at mail.
- Search form.
- Structure of a Genetic algorithm?
- OS X Mountain Lion all-in-one for dummies;