DEVELOPMENT OF A SMELL AGENT OPTIMIZATION ALGORITHM FOR COMBINATORIAL OPTIMIZATION PROBLEMS
Need help with a related project topic or New topic? Send Us Your Topic
DOWNLOAD THE COMPLETE PROJECT MATERIAL
DEVELOPMENT OF A SMELL AGENT OPTIMIZATION ALGORITHM FOR COMBINATORIAL OPTIMIZATION PROBLEMS
Chapter one
INTRODUCTION: 1.1 Background
Efforts to create an accepted definition of intelligence continue to spark discussions across disciplines. Dictionaries (Crystal, 2004; English, 2007) define intelligence as the ability to understand, grasp, and profit from experience, as well as the ability to interpret and reason, particularly to a high level.
All living systems demonstrate “intelligence” mechanisms that are similar in complexity, structure, and overall adaptability. Over the years, scientists have sensibly sought ways to codify intelligence systems into algorithms devoted to addressing some complicated problems in engineering and related fields.
This sparked the creation of a new field of research known as computational intelligence (CI), which was popularised by James C. Bezdek approximately 24 years ago (Bezdek, 1994).
Perhaps the earliest appearance of CI occurred in 1983, when Gordon McCalla and Nick Cercone, the editors and founders of the Canadian publication, reported the title as the International publication of Computational Intelligence (IJCI) (Bezdek, 2013).
According to Mu’azu (2006, 2016), computational intelligence encompasses science-based tools and methodologies for analysing and constructing intelligent systems.
The term “intelligent” refers to engineering techniques based on human thinking, adaptation, learning, biological cognitive structures, evolutionary concepts, and natural physical or chemical processes.
2
Many experts have also defined computational intelligence (CI) as a computer system’s ability to learn a certain task from a particular amount of data and/or empirical observation (Siddiqui & Hojjat, 2013).
The IEEE World Congress on Computational Intelligence and the IEEE Neural Network Council established this term in the summer of 1994 in Orlando, Florida. (Xing and Gao, 2014).
Perhaps scholars have yet to reach an agreement on a clear definition of computational intelligence (Zimmermann, 1999). This is due to the difficulty of starting with something precise; accuracy must be achieved through a specific method.
However, properly speaking, computational intelligence is a set of nature-inspired computational models and methodologies capable of addressing real-world challenges to which traditional or mathematical models may be limited due to one or more of the following reasons:
1) The problem or procedure could be too complex for mathematical thinking.
2) The problem or process could be dynamic and stochastic in nature.
3) The problem’s solution space may be too big to handle mathematically.
4) The problem or procedure may involve uncertainties.
Most real-world (non-linear) science, economic, social, and engineering problems share all of these traits. These non-linear issues necessitate various assumptions in order to be converted to their near-linear equivalents for simple computation.
However, the outputs of such linear computations do not always accurately represent real-life circumstances, unless the exact answer is not crucial. Thus, agent-based computational intelligence systems can provide excellent and promising options in such a circumstance.
3
The most evident high-level intelligent agents are humans. An intelligent agent is a rational creature that acts relatively autonomously (independently) in its environment on behalf of its user.
However, there are classes of intelligent agents (for example, natural phenomena such as water droplets, flower pollination, river formation, etc., or animals such as fish, dogs, worms, insects, etc., or even bacteria, amoeboids, etc.) that may be more intelligent than humans in terms of coordination and organisation (Poole & Mackworth, 2010).
Ant colonies are one example of such a sophisticated organisation. A single ant may not be very intelligent, but the entire colony will function more wisely and efficiently than any individual ant.
The ant colony can detect the presence of food and search for it in order to successfully exploit the food supply while also adapting to changes in the environment through the use of specialised skills.
Several researchers have created agent-based biological and nature-inspired CI algorithms during the last few decades to solve a variety of optimisation challenges. The foraging activities of ant systems were formalised into an algorithm by Dorigo et al. (1996).
Algorithms based on the intelligent behaviour of fish (Lei et al., 2002), bacterial foraging (Passino, 2002), firefly (Yang, 2010), swarm particles (Eberhart & Kennedy, 1995), and bee (Karaboga & Basturk, 2007) have also been created.
The performance of all of these algorithms on appropriate tasks, such as in (Malarvizhi & Kumar, 2015; Pradhan et al., 2016; Sundaram et al., 2016; Tijani & Mua’zu, 2015; Turabieh & Abdullah, 2011, has proved their usefulness in tackling real-world situations.
However, it is vital to remember that no single nature-inspired optimisation strategy exists that can solve all optimisation difficulties. This is known as the “no free lunch” theorem (Wolpert and Macready, 1997).
Need help with a related project topic or New topic? Send Us Your Topic