Monthly Archives: June 2024

A metaheuristic is a higher-level procedure or heuristic used in computer science and mathematical optimization to identify, create, adjust, or choose a heuristic (partial search algorithm) that can adequately solve an optimization or machine-learning problem, particularly when there is limited computing power or incomplete or imperfect information available. Metaheuristics are near-optimal solution methods used to solve NP-hard optimization problems. Metaheuristics may be used for a wide range of situations because they tend to make minimal assumptions about the optimization problem that must be addressed.

Due to limitations in time, space, and resources, it is difficult to explore all potential solutions to the many real-world problems encountered. Thus, it is essential to use faster, more cost-effective, and technologically superior techniques. As a result, several algorithms have been developed based on the lives, behaviour, hunting, and self-defense techniques of various species in nature, such as fish schools, krill herds, fox packs, bee colonies, red foxes, and whales. These algorithms are referred to as nature-inspired algorithms (NIAs) because they were developed based on this principle. Considering the abundance of traditional optimization methods, one may naturally wonder why researchers need novel algorithms, such as social algorithms. Traditional algorithms perform well in a wide range of problem types, according to the literature and substantial research; however, they have several significant drawbacks. Traditional algorithms mostly rely on local search and do not provide global optimality in the majority of optimization tasks. Because they often need knowledge, such as derivatives of the local objective landscape, they are typically problem-specific. Multimodal and highly nonlinear issues are too complex for traditional algorithms to handle. They struggle to deal with discontinuity problems, particularly when gradients are required. Because they are often predictable, they have strong exploitation ability, but a poor capacity for exploration and a variety of solutions. In this study, we addressed and improved the above-mentioned problems by proposing a new metaheuristic algorithm.

The GOOSE algorithm, a new metaheuristic algorithm based on the behaviour of geese during rest and foraging, is proposed. The goose balances and stands on one leg to monitor and guard the other birds in the flock. Notably, the GOOSE method is a particle swarm optimization (PSO) -based approach that updates the location of the search agent with the addition of velocity. The GOOSE algorithm is described throughout this work of art along with an explanation of the idea's inspiration.

The accuracy and performance of the proposed algorithm were rigorously verified by testing it on various benchmark functions. The GOOSE algorithm was benchmarked on 19 well-known benchmark test functions, and the results were verified through a comparative study with a genetic algorithm (GA), (PSO), dragonfly algorithm (DA), and fitness-dependent optimizer (FDO). In addition, the proposed algorithm was tested on ten modern benchmark functions, and the obtained results were compared with three recent algorithms: the dragonfly algorithm, whale optimization algorithm (WOA), and salp swarm algorithm (SSA). Moreover, the GOOSE algorithm was tested on five classical benchmark functions, and the obtained results were evaluated using six algorithms: the fitness-dependent optimizer, FOX optimizer, butterfly optimization algorithm (BOA), whale optimization algorithm, dragonfly algorithm, and chimp optimization algorithm (ChOA). The obtained findings attest to the superior performance of the proposed algorithm compared with the other algorithms utilized in the current study. The technique is then used to optimize the welded beam design, Economic Load Dispatch Problem, Pressure Vessel Design Problem, and the Pathological IgG Fraction in the Nervous System, four well-known real-world challenges. The outcomes of engineering case studies illustrate how well the suggested approach can optimize real-world issues.

Comparison of GOOSE statistical results with literature for the welded beam design problem compared to other algorithms such as WOA, PSO, and GSA, where the Goose algorithm ranked third in the list with an average of 3.1882. Also, GOOSE statistics findings are compared with the literature for the pressure vessel design problem, and the method ranks second with an average of (6343.6587) when compared to other algorithms such as WOA, PSO, and GSA. On the other hand, the goose algorithm was used to improve a health problem called (The Pathological IgG Fraction in the Nervous System), and as a result, our algorithm achieved a very good comparative result to the LEO algorithm with an average of (0.00047792). Because this application is newly designed and only optimized by the LEO algorithm, it was compared with the LEO results.

Fog Computing (FC) has recently emerged as a promising new paradigm

that provides resource-intensive Internet of Things (IoT) applications with low

latency services at the network edge. However, the limited capacity of

computing resources in Fog colonies poses great challenges for scheduling and

allocating application tasks. In this dissertation, an Intelligent Scheduling

Strategy Algorithm in a Fog Computing system based on Multi-Objective Deep

Reinforcement Learning (MODRL) is proposed. MODRL algorithm select

nodes (Fog nodes or Cloud nodes) for task processing based on three

objectives; current node’s Load, node Distance, and task Priority. MODRL is

a smart method that integrates the ideas of Multi-Objective Optimization and

Deep Reinforcement Learning to tackle intricate decision-making situations

involving several conflicting objectives. This technique is especially valuable

in situations when there is a requirement to maximize numerous criteria

simultaneously, even if they do not exactly line, and where trade-offs need to

be taken into account. The proposed model addresses two main problems; task

allocation and task scheduling. Employ three Deep Reinforcement Learning

(DRL) agents based on a Deep Q Network (DQN), one for each objective. It is

a specific form of Artificial Neural Network structure employed in

Reinforcement Learning. The DQN algorithm utilizes a Deep Neural Network,

commonly a Convolutional Neural Network (CNN), to estimate the Q-function.

This enables the model to effectively process intricate input domains. However,

this is a more challenging scenario because there is a trade-off among these

objectives, and eventually, each algorithm may select different processing

nodes according to its own objective, which brings to a Pareto Front problem.

To solve this problem, propose using Multi-Objective Optimization, a Non

dominated Sorting Genetic Algorithm (NSGA2), and a Multi-Objective

Evolutionary Algorithm based on Decomposition (MOEA/D), which are Multi-VII

Objective Optimization algorithms that can choose the optimal node by

considering three objectives.

Simulation investigation and experiments using a Python environment

with TensorFlow, PyTorch, Pymoo, and PQDM libraries in PyCharm, which is

a powerful Python IDE, to simulate and train the Intelligent Scheduling

Strategy. As well as, Virtualized data using MatPlotLib in the Jupyter

Notebook, indicates that the proposed Intelligent Scheduling Strategy could

attain better results for the several employed efficiency, adaptability, and

performance metrics: Task Completion Time, Makespan, Transmission Delay,

Queueing Delay, Processing Delay, Propagation Delay, Computational Delay,

Latency, Network Congestion, Throughput, CPU Load, and Storage

Utilization, with an average value of 2.02ms, 10ms, 25ms, 2ms, 1.0ms,

9.5ms,3ms, 3.5ms, 0.10ms, %100, %10, and % 99, respectively.