Abstract
Noninvasive process of vital sign identification and design of low-cost decision-making system for the betterment of rural health care support is a prime facet of research. The identification of bio-signals from different sensors, noise removal, signal processing, and decision making requires the use of sophisticated expert system. In this chapter, the authors propose a modified multi grey wolf pack optimization technique (MMGWO) for better generalization and diversification. The basic model has been modified using net energy gain of the individual wolf in the process of hunting. Here multiple packs of wolves are considered with simultaneous sharing and conflict among them. The performance of the proposed technique is tested on 23 well known classical benchmark functions, CEC 2014 benchmark problem set along with classical real-life applications. The experimental results and related analysis show that the proposed MMGWO is significantly superior to other existing techniques.
TopIntroduction
Conventional diagnostic practices are very time consuming, labor-intensive and these require elaborate infrastructures and on-field expertise. It is difficult to set up such facilities in remote places and it is further difficult to obtain expert medical team for those places to cater the medical need for the general people with the available medical infrastructure. In such a scenario, design of non-contact and/or noninvasive medical support system may be considered as a primary target to connect those places for the minimum health care support. The implementation of such a scheme requires the widespread use of the sophisticated and automated expert system for disease detection. Here, basic bio signals are collected by different noninvasive methods using ensemble of sensors. Noises are removed from those signals; the signals are processed, decisions are taken and they are conveyed to medical experts present far away for better advice. The above methodology requires the design of an expert system with sophisticated sensor ensemble which collects various pathological parameters like blood pressure, pulse rate, SpO2, perfusion rate, activity profile etc and they are transmitted to the signal processing module. The signals collected amidst real life environment needs robust and advanced level of signal processing algorithms. After getting the processed signals the hidden complex features are extracted and those non-redundant optimal set of features act as a fundamental backbone for first level of disease diagnosis and screening by decision making system. Now all the above modules like signal processing module, feature extraction module and above all the decision making module in individual stages require robust optimization algorithms for balanced and automated operation to cater the non-invasive remote medical functionality in remote area. The feature extraction module requires the application of robust optimization algorithm for feature extraction. Also the decision making system requires different mechanism like clustering, classification of data and parametric optimization for effective decision making and successively, a routine availability of minimum medical facilities may be ensured early in time to the people of remote places before the situation becomes a serious issue. Therefore, it may be observed that at all individual level of such kind of expert system design requires robust, auto adjustable, optimization algorithms with dedicated functionalities at different stages and the performances of those algorithms are finally going to be the driving factor in deciding the effectiveness of the support system. This will eventually decide the success of the system in providing the medical facility to the remote and rural healthcare paradigm. Inspired by the above requirement, in this book chapter, we propose to work on the most important aspect of such systems, i.e., on the design and development of sophisticated, simple but noble optimization algorithm and that may act as a back bone of any automated non-invasive health diagnostic system. Also, the above discussion may be supported in the review section where we find the immense use of optimization algorithms in the design and development of non-invasive technology-based diagnostic methods.
Key Terms in this Chapter
Swarm Intelligence (SI): It is basically population-based population. It depends on natural behavior of animals, bird etc. It utilizes the concept that a number of particles (candidate solution) which fly around the search space to find the best solution.
Optimization: Optimization means the process of finding the best solution(s) of a particular problem(s).
Energy Equation: Energy, in physics, is the capacity for doing work. It may be potential, kinetic, thermal, electrical, chemical, nuclear energy, etc.
Crossover: Means the recombination that means a genetic operator used to combine the genetic information of two parents to generate new offspring.
Multi Population: Multi population divides the whole population into multiple subpopulations. The population diversity can be maintained because different subpopulations can be located in different search spaces.
Local Optima: In optimization, the local optima are the relative best solutions within a neighbor solution set.
Meta-Heuristic Algorithms: A metaheuristic is a higher-level procedure or heuristic designed to find a good solution to an optimization problem, especially with incomplete or imperfect information or limited computational capacity.