DATA SEARCH OPTIMIZATION USING BIO-INSPIRED TECHNIQUES

Abstract

All the information available these days are available in the form of digital data, whether it is network information, sensor information, healthcare information or from some other field. With the deluge of data every day the requirement to store and process the data is increasing. So technology to store, process and analyze the data should also improve with the same stride. Researchers are innovating new ideas to process the data faster with new algorithms. There has been many research in the area of Data Science which proposed the parallel processing architecture, some tools and framework has been introduced to implement the parallel processing. But apart from parallel processing, optimization of algorithm on each node of parallelism is also very important. There are many activities in the nature which are happening perfectly optimized, without any external control or without any centralized structure. And sometime these activities have been proven to solve the most difficult problems observed by the traditional computing method. In this thesis we have presented a solution inspired by nature to optimize the Data Search for shortest possible path among given locations. The proposed algorithm has been applied on existing parallel processing framework “MapReduce” with the slight modification in the MapReduce architecture to find the global best among local best produced by each Map node. So MapReduce, which was twostep process earlier has been modified to three step process Map-Reduce-Reduce. And for finding shortest path inspiration has been taken from Bottlenose dolphin. Bottlenose Dolphins are well known for their intelligence and communication; they use echolocation to identify the prey location and calculate the shortest path with the group effort of all the dolphins in the group. Utilizing this behavior of bottlenose dolphin a new heuristic approach has been developed and implemented in this thesis to enhance the searching process for the shortest path.