Explore chapters and articles related to this topic
Cuckoo Search Algorithm
Published in A Vasuki, Nature-Inspired Optimization Algorithms, 2020
Since the original cuckoo search algorithm was proposed in 2009, several variants of the algorithm have been developed by various researchers working in the area of swarm intelligence [5]. These variants either modify the algorithm to be more efficient or make it adaptable for diverse applications. Some of the notable variants of the algorithm are modified CS, parallelized CS, binary CS, discrete CS, neural-based CS, and multi-objective CS. For some problems which are difficult to solve, especially continuous optimization problems, the optimal solution can be obtained efficiently by combining the CS algorithm with some other swarm-based algorithm such as GA or bat algorithm so that a hybrid optimization algorithm evolves to find a better solution to the problem. Multi-objective optimization is another variant where the problem consists of conflicting, multiple objectives. A set of solutions forming a Pareto Optimal Front will be suitable for such multi-objective problems. The CS algorithm has been adapted to solve specific engineering design problems such as wind turbine blades, steel truss structures, antenna arrays, optimal sequence attainment, optimal capacitance placement, and also NP-hard problems such as TSP, job scheduling, and graph coloring. The three notable variants of the CS algorithm are discussed below.
Introduction
Published in Jun Ma, Xiaocong Li, Kok Kiong Tan, Advanced Optimization for Motion Control Systems, 2020
Jun Ma, Xiaocong Li, Kok Kiong Tan
Optimization of a collection of objectives systematically and simultaneously is a multi-objective optimization problem, where these objectives are the criteria of different design targets. These design targets could be incommensurable and competing, and it becomes a burden to find the optimal solution with trade-off surfaces. Classical methods to obtain the optimal solution will aggregate multiple objectives into a single form parameterized objective function, such as the weighting method and the constraint method. The weighting method converts a multi-objective optimization problem into a single optimization problem by forming a linear combination of the objectives. However, this approach is not capable of deriving the optimal solution with non-convex trade-off surfaces. On the other hand, the constraint method transforms multiple objectives into constraints. Remarkably, the constraint method is suitable for the case where a prior preference on the multiple objectives is available.
Decision Making for Multi-Objective Life-Cycle Optimization
Published in Dan M. Frangopol, Sunyong Kim, Life-Cycle of Structures Under Uncertainty, 2019
The development of novel probabilistic concepts and methods for optimum inspection, monitoring and maintenance planning has led to an increasing demand of objectives to be considered in the optimization process (Frangopol 2011; Frangopol and Soliman 2016; Kim and Frangopol 2017). Multi-objective optimization has been treated as an effective tool for integrating multiple objectives and for providing multiple well-balanced solutions. However, an increase in the number of objectives generally requires additional efforts for computation in order to find the entire Pareto optimal solutions, visualization of these solutions, and decision making to select the best Pareto optimal solution (Deb and Saxena 2006; Saxena et al. 2013; Verel et al. 2011). Therefore, it is necessary to develop a decision making framework for optimum inspection and monitoring planning, to deal with a large number of objectives efficiently and to select the best single optimum inspection and monitoring plan for practical applications (Brockhoff and Zitzler 2009; Kim and Frangopol 2018a, 2018b).
Techno-economic assessment of photovoltaic-based charging stations for electric vehicles in developing countries
Published in Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2023
Loiy Al-Ghussain, Adnan Darwish Ahmad, Ahmad M. Abubaker, Mohammad Alrbai, Osama Ayadi, Sameer Al-Dahidi, Nelson K. Akafuah
Several algorithms, such as the one in this work, can be used to optimize the multi-dimensional problem. Genetic algorithm, (pattern) search, Particle swarm, and non-dominated sorting genetic algorithm are some algorithms used in multi-objective optimization problems. These would be considered population optimization algorithms which are stochastic that maintain a pool of candidate solutions that are gathered and used to represent, explore, and improve on an optimum. The algorithms can solve problems with noisy function evaluations and several global optima. Literature has shown the reliability of such algorithms. Pareto search was used in the 100% renewable energy systems, as shown in Al-Ghussain et al. (2022). A genetic algorithm is used in a hybrid system consisting of fossil fuel and solar power, as reported in Abubaker et al. (2022).
Dynamic Resource Allocation Using Improved Firefly Optimization Algorithm in Cloud Environment
Published in Applied Artificial Intelligence, 2022
Simin Abedi, Mostafa Ghobaei-Arani, Ehsan Khorami, Musa Mojarad
Optimization algorithms seek to find a solution in the search space that has the minimum (or maximum) value of the objective function (Hajipour, Khormuji, and Rostami 2016). In an optimization problem, the types of mathematical relationships between objectives, constraints, and decision variables characterize the difficulty of the problem. Objective optimization is often used by researchers to solve real-world problems (Yang and Deb 2009). In general, a better solution is achieved by setting several objectives in solving the problem. Multi-objective optimization is a process for solving a problem with simultaneous optimization of two or more objectives subject to constraints. In recent years, nature-inspired algorithms and biological processes have been the most powerful solutions to optimization problems (Hajipour, Khormuji, and Rostami 2016; Mirjalili and Lewis 2016; Yang and Deb 2009).
Multi-objective optimization of pavement preservation strategy considering agency cost and environmental impact
Published in International Journal of Sustainable Transportation, 2021
Israa Al-Saadi, Hao Wang, Xiaodan Chen, Pan Lu, Abbas Jasim
Multi-objective optimization is needed when decision makers want to achieve two or more objectives at the same time. Transportation agencies try to find an adequate maintenance strategy to minimize both agency cost and user cost. In general, minimizing user costs requires frequent maintenance treatments to keep good pavement condition, which in turn results in increasing agency cost. Although pavement maintenance may cause the increase of agency cost (due to increased material and construction) and user cost (due to traffic delay in work zone), the reduction of vehicle operation cost in the use stage of pavement may be dominant for high traffic volume. Therefore, these two competing objectives (agency cost and user cost) are contradictory in many conditions. There is a range of measures that can be taken to unify objectives that appear to be contradictory (Abaza, 2007; Mbwana, 2001). Thus, the optimal solutions should compromise one objective for the benefit of another, in other words, there does not exist another feasible solution that yields an improvement in one objective without causing the degradation of the other one. The set of the optimal solutions are called Pareto solutions. As such, the goal is to find final solution that is acceptable and balanced.