University Research Project: Consolidation of AI Search Algorithms

This research project delves into the consolidation of search algorithms within artificial intelligence (AI) systems. It begins by defining search in the context of AI and highlights the different types of search, including uniformed, informed, and local search, along with the algorithms that fall under these categories. The project emphasizes the importance of algorithm selection and the limitations of relying on a single algorithm for all AI problems. It explores the concept of algorithm portfolios, where different algorithms are used on a case-by-case basis to optimize performance. The research investigates the application of consolidated search algorithms in various fields, using examples such as virtual machine consolidation in cloud computing and exoplanet detection using neural networks. The methodology involves qualitative research, specifically content analysis of visual and textual materials, including surveys of AI experts. The project aims to identify the different ways in which search algorithms are consolidated, the reasons for this consolidation, and the benefits it offers to organizations using AI. The research is expected to provide valuable insights into the significance of consolidating AI search algorithms and their practical applications. The project will benefit organizations using AI and those considering adopting this technology, offering insights into the advantages of consolidated search algorithms.

Running head: CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
Name of the Student
Name of the university
Author note

1CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
Introduction
Search in the context of artificial intelligence or AI refers to the process involved in
navigation from a starting state to a specific goal state by going through different
intermediate states. In general all AI problems can certainly be defined in the terms of state,
transition, starting state, intermediate state, goal state and the search space. There are various
different types of search in AI such as uniformed search, informed search and local search.
Under these search types there come different algorithms which can be said to be as the
instructions given to the AI systems (Elsken, Metzen and Hutter 2018). Search therefore has a
major role to play in providing solutions to many different AI problems and can be said to as
a universal mechanism for solving issues in AI. The main aim of the research is to discuss on
the consolidation of various search algorithms for the AI systems.
Background
As per the research done by Kotthoff (2016), selection of the best algorithm plays a
vital role in solving any problem. In the context of combinational search problems, there have
been many contributions made in the field of AI. In some cases it has been found that a new
algorithm is superior to the previous approaches and that this new approach brings in
improvisation in the current state of the art for some specific problems. This can be said to be
because of the employment of heuristic approach that cannot be applied to all problems. Thus
it can be certainly said that algorithm selection has become a problem in many different
forms with distinct names in various areas of AI. There have been researchers done to tackle
this new problem. Researchers that have been conducted have come to a point that a single
algorithm cannot provide best performance across all the AI problems and there comes the
need of consolidation of the search algorithms. Thus, the way out to the problem is making

2CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
use of portfolio of different algorithms. This idea basically is about deciding which
algorithm to use on a case-by-case basis rather than just getting straight to an algorithm.
As opined by Suprakash and Balakannan (2018), with demand in internet mobile
computing and introduction of 3G and 4G networks, the demand of the computing resources
at datacenters leading to the consumption of excessive amount of power. The authors propose
unique solution or approach for bringing in consolidation of multiple virtual machines in the
context of cloud based distributed hosting surroundings. For this purpose algorithm that has
been used is penguin search optimization algorithm or PeSOA, Memetic and particle swarm
algorithm. Virtual machines are a part of AI systems and in this paper virtual machine
consolidation has been attempted with respect to different algorithms one being dependent on
penguin’s behaviour of searching as well as selecting for food resources. The parameters that
have been considered in the case are the CPU of virtual machines and availability of memory.
The algorithm is as given below:
Figure 1: PESOA Algorithm (Suprakash and Balakannan 2018)
The other algorithms that have been mentioned above have also been discussed and all of
these been tested with Cloud Sim that refers to a library related to the Simulation of Cloud
Computing Scenarios. The results of the study are clearly evident of the fact that one of the
algorithms which are the Memetic algorithm perform better than the other two thus giving a

3CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
reason to search for best algorithms or opt for consolidation of search algorithms in case of
AI.
The authors Pearson, Palafox and Griffith (2018) suggest that manual interpretation of
the potential exoplanet candidates is not only labour intensive but also subjected to various
human errors and the results in this case are not easily quantifiable. To help in this case
neural network also termed as deep nets are developed to provide a computer perception into
a certain problem by means of training the same to recognize different patterns. Unlike the
traditional algorithms deep nets learn to find out features of the planet rather than just relying
upon hand-coded metrics that are perceived by the humans as the most representative. The
SNR of transit detection can certainly be maximized when convolving the data with optimal
filter. Convolution neural networks have been effective in beating human perceptions at
object recognition. In this context algorithms play an important part as these are made use of
for removing substantial fraction of false positive signals prior to the detection of the signal
from the planet. Neural networks consist of different technologies such as deep learning as
well as machine learning being part of AI. In this context the study has been referred to as
neural networks are made use of for the purpose of evaluation in the Heuristic search
algorithms.
Research Questions and Objectives
The research gap that has been identified in this context is that not many researches
talk about how consolidated search algorithms in AI are being used in different spheres or
what actually is the underlying concept behind this. Another main question that has not been
answered is the significance of consolidation of search algorithms of AI.
Research question:

4CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
1. What are the different ways in which search algorithms of AI are being consolidated to use
the same in different fields or for various purposes?
2. What is the need for consolidation of search algorithms of AI?
Research objective:
1. To figure out the different ways in which search algorithms of AI are being consolidated to
use the same in different fields or for various purposes.
2. To find out the underlying reason as to why consolidation of algorithms of AI is needed.
Research Methodology
In this research qualitative research methodology will be made use of. The main
research method that will be used is content analysis of visual as well as textual materials that
are related to this topic. Sites such as Google scholar will be referred to for the purpose. In
this case open-ended surveys that fall under the purview of qualitative research can also be
made use of (Glesne 2016). The survey participants can be experts from different
organizations those who are working with AI. The reason for selecting qualitative research is
that will provide a in depth as well as detailed knowledge thus helping to analyze the matter.
This can be said to be as a more flexible research methodology rather than the quantitative
research methodology thus will help in making speculations about the specific areas that have
been chosen for the investigation (Fletcher 2017). This research methodology helps
researchers to adapt well in situations where useful insights are not being obtained thus aiding
to change the setting as well as other variable so that the responses can be improved. Open-
ended has been opted for as one of the research methods as this will help to understand the
perception of people and help in gaining practical idea about the topic (Choy 2014).

5CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
Task list
Activities during the research Time required
Selecting the topic week 1
Project proposal week 2
Literature review week 3 and week 4
Analysis of literature review week 5
Data collection (This will include preparing the
survey questionnaire and selecting participants
and distribution of the questionnaire through
online medium)
week 6 and week 7
Analysis of the results of data collection week 8
Draft paper week 9
Final thesis week 10
Project type and expected outcomes
The project will help other researchers in this field to understand the purpose of
consolidation of search algorithms in AI. The main purpose of the research is providing the
answer to one of the most important questions that is where or in which fields consolidation
of AI search algorithms are being made use of and what are the advantages of these. The
research attempts to figure out the difference in using single algorithms in the AI systems and
consolidated search algorithms in the same. This is desired that by the end of the research the
concept of search algorithms in the context of AI will be clear to the readers along with the
emerging need to consolidate all these. The research’s purpose is clear and vivid as this will
focus on the reason behind the need for consolidation of different search algorithms in AI and
the advantages of the same.

6CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
Significance
The project is expected to benefit different organizations that are making use of AI as
well as even those that are thinking of adopting this new technology for the sole purpose of
brining in improvisations in their business operations. The use of AI has grown almost in
very sector and thus it can be said that this research on the consolidation of search algorithms
in AI is required and essential. The world today is anticipating that in days to come AI with
its wide range of applications will be made use of in different spheres thus it can be said that
at this stage the research is most required as well as necessary for the people.

7CONSOLIDATION OF SEARCH ALGORITHMS FOR AI SYSTEMS
References
Choy, L.T., 2014. The strengths and weaknesses of research methodology: Comparison and
complimentary between qualitative and quantitative approaches. IOSR Journal of Humanities
and Social Science, 19(4), pp.99-104.
Elsken, T., Metzen, J.H. and Hutter, F., 2018. Neural architecture search: A survey. arXiv
preprint arXiv:1808.05377.
Fletcher, A.J., 2017. Applying critical realism in qualitative research: methodology meets
method. International journal of social research methodology, 20(2), pp.181-194.
Glesne, C., 2016. Becoming qualitative researchers: An introduction. Pearson. One Lake
Street, Upper Saddle River, New Jersey 07458.
Kotthoff, L., 2016. Algorithm selection for combinatorial search problems: A survey. In Data
Mining and Constraint Programming (pp. 149-190). Springer, Cham.
Pearson, K.A., Palafox, L. and Griffith, C.A., 2018. Searching for exoplanets using artificial
intelligence. Monthly Notices of the Royal Astronomical Society, 474(1), pp.478-491.
Renzi, C., Leali, F., Cavazzuti, M. and Andrisano, A.O., 2014. A review on artificial
intelligence applications to the optimal design of dedicated and reconfigurable manufacturing
systems. The International Journal of Advanced Manufacturing Technology, 72(1-4), pp.403-
418.
Suprakash, S. and Balakannan, S.P., 2018. A POWER EFFICIENT VIRTUAL MACHINE
CONSOLIDATION FOR SAVING CARBON FOOTPRINT USING META HEURISTIC
ALGORITHMS. International Journal of Pure and Applied Mathematics, 119(16), pp.4027-
4034.