Collective Intelligence Algorithms Research Articles

Optimizing Variational Quantum Neural Networks Based on Collective Intelligence

Quantum machine learning stands out as one of the most promising applications of quantum computing, widely believed to possess potential quantum advantages. In the era of noisy intermediate-scale quantum, the scale and quality of quantum computers are limited, and quantum algorithms based on fault-tolerant quantum computing paradigms cannot be experimentally verified in the short term. The variational quantum algorithm design paradigm can better adapt to the practical characteristics of noisy quantum hardware and is currently one of the most promising solutions. However, variational quantum algorithms, due to their highly entangled nature, encounter the phenomenon known as the “barren plateau” during the optimization and training processes, making effective optimization challenging. This paper addresses this challenging issue by researching a variational quantum neural network optimization method based on collective intelligence algorithms. The aim is to overcome optimization difficulties encountered by traditional methods such as gradient descent. We study two typical applications of using quantum neural networks: random 2D Hamiltonian ground state solving and quantum phase recognition. We find that the collective intelligence algorithm shows a better optimization compared to gradient descent. The solution accuracy of ground energy and phase classification is enhanced, and the optimization iterations are also reduced. We highlight that the collective intelligence algorithm has great potential in tackling the optimization of variational quantum algorithms.

An Optimal-Path-Planning Method for Unmanned Surface Vehicles Based on a Novel Group Intelligence Algorithm

Path planning is crucial for unmanned surface vehicles (USVs) to navigate and avoid obstacles efficiently. This study evaluates and contrasts various USV path-planning algorithms, focusing on their effectiveness in dynamic obstacle avoidance, resistance to water currents, and path smoothness. Meanwhile, this research introduces a novel collective intelligence algorithm tailored for two-dimensional environments, integrating dynamic obstacle avoidance and smooth path optimization. The approach tackles the global-path-planning challenge, specifically accounting for moving obstacles and current influences. The algorithm adeptly combines strategies for dynamic obstacle circumvention with an eight-directional current resistance approach, ensuring locally optimal paths that minimize the impact of currents on navigation. Additionally, advanced artificial bee colony algorithms were used during the research process to enhance the method and improve the smoothness of the generated path. Simulation results have verified the superiority of the algorithm in improving the quality of USV path planning. Compared with traditional bee colony algorithms, the improved algorithm increased the length of the optimization path by 8%, shortened the optimization time by 50%, and achieved almost 100% avoidance of dynamic obstacles.

Decision Support System for the Integrated Management of Multiple Supply Systems in the Brazilian Semiarid Region

To alleviate the impacts of water shortages and ensure the effective allocation of water resources, the integration of multiple sources has been a viable but challenging alternative for water resource managers. A robust decision support system (DSS) reduces the complexity of decision making and allows the user to define operational strategies that reduce costs and address system limitations. Through the Particle Swarm Optimization (PSO) collective intelligence algorithm and the operational zoning of reservoirs, it was possible to identify the monthly flow allocation for the Agreste of Pernambuco in Brazil. The analysis of unit costs allowed us to conclude that high local energy tariffs guide the DSS in the exploration of water-transfer works between basins, such as the Agreste pipeline, which had full use of approximately 2 m3/s in the analyzed scenarios. The study also sought to investigate the influence of rainy and dry periods, with the higher flows into local reservoirs in rainy periods leading the system to operate at a cost 2.6% higher compared to the dry period. The use of DSS was able to provide support and parameters that facilitate decision making, ensuring that the system operates efficiently and sustainably.

Optimal dispatch of dynamic power and heat considering load management, water pump system, and renewable resources by grasshopper optimization algorithm

Nowadays, one of the most important issues in a country's power system is to meet the demand for electrical energy. However, an essential aspect of this goal is to minimize production costs. It also must be taken into account that this is a very complicated problem, considering nonlinear constraints and uneven cost function. It should be solved by an optimization method with high reliability that could be implemented easily. Given the non-convex nature of this problem, it will be solved by difficult mathematical methods and a high computational load. In this paper, the economic power dispatch during 24 h/day becomes an optimization problem. In order to create a real problem, load dispatch is modeled considering the heat, renewable, and water resources for demand-side management. Since the proposed model is very complicated, we used the grasshopper optimization algorithm. This algorithm belongs to the group of collective intelligence algorithms and is inspired by the social behavior of grasshoppers and the impact of the environment on each grasshopper. In this algorithm, updating the position of each grasshopper depends on the distance of each grasshopper from the entire population of grasshoppers in the current generation and the position of the best grasshopper. Features of this algorithm include simplicity and having only one adjustable parameter. The simulation results indicated the appropriate effectiveness of load management in a cost.

ЭФФЕКТИВНЫЕ ОБРАЗОВАТЕЛЬНЫЕ ПРАКТИКИ НА ОСНОВЕ АЛГОРИТМОВ ИСКУССТВЕННОГО ИНТЕЛЛЕКТА В ПОДГОТОВКЕ БУДУЩИХ ПЕДАГОГОВ

The paper considers practical aspects of the implementation of the educational process in an electronic educational environment for students — future teachers in order to form their readiness to perform work functions and actions in conditions of uncertainty and digital transformations. Based on the analysis of scientific literature, the results of the survey, the experience of preparing students of pedagogical directions, as well as taking into account the application of a multi-paradigm approach to pedagogical education, the methods of training future teachers using artificial intelligence algorithms are proposed. The article presents the methodology of the existence and functioning of a collective pedagogical megasystem, both forecasting pedagogical processes and identifying new patterns and making effective didactic decisions based on the use of collective intelligence algorithms in the implementation of pedagogical practice. The article reveals the applied aspect of the use of collective intelligence algorithms in the implementation of quasi- professional activities at the university and during pedagogical practice in order to form the readiness of future teachers to implement pedagogical activities based on learning to choose effective didactic solutions. This aspect is key for the readiness and ability to implement professional actions and functions due to variations: the contingent of students, subject content, organizational and pedagogical conditions of a particular educational organization. Taking into account both spontaneous and systemic aspects in the implementation of professional activity is crucial for the choice of learning technologies that ensure maximum productivity of the teacher and, as a result, the achievement of high educational results of students. For these purposes, the most suitable algorithm for choosing methods and technologies for the implementation of professional activities by future teachers is the ant algorithm. This algorithmic model allows us to take into account both the probabilistic nature of the changing impact parameters, and allows us to evaluate and select the high effectiveness of the application of this algorithm, which distinguishes it favorably among other artificial intelligence algorithms. The article provides an example of an ant algorithm for determining the best hierarchy of organizational and pedagogical conditions conducive to the best choice of technologies and methods for the implementation of professional activities by future teachers. As a conclusion, the authors propose to use collective intelligence algorithms in the process of quasi-professional activity of future teachers in the framework of pedagogical practice, which will contribute to the formation of their readiness to implement pedagogical activities in terms of performing labor functions and actions based on learning to choose effective didactic solutions.

Bacterial Foraging Based Algorithm Front-end to Solve Global Optimization Problems

The Bacterial Foraging Algorithm (BFOA) is a well-known swarm collective intelligence algorithm used to solve a variety of constraint optimization problems with wide success. Despite its universality, implementing the BFOA may be complex due to the calibration of multiple parameters. Moreover, the Two-Swim Modified Bacterial Foraging Optimization Algorithm (TS-MBFOA) is a state-of-the-art modification of the BFOA which may lead to solutions close to the optimal but with more parameters than the original BFOA. That is why in this paper we present the design using the Unified Modeling Language (UML) and the implementation in the MATLAB platform of a front-end for the TS-MBFOA algorithm to calibrate the algorithm parameters faster and with no need for editing lines of code. To test our proposal, we solve a numerical optimization problem with constraints known as tension/compression spring, where 30 independent executions were conducted using the TS-MBFOA and then compared with an earlier version called MBFOA. The runtime configuration and the parameter tuning were fluent using our front-end, and the TS-MBFOA obtained the better results. To date, there is no other user-friendly implementation of this specific algorithm in an open-source code, and the front-end is flexible enough to include other numerical optimization problems with minimal effort.

Nonlinear Modeling for Low Frequency Oscillations Damping Using the Collective Intelligence Algorithms

Abstract This paper presents a design for a fuzzy controller for low frequency oscillations using the improved virus colony search. In this way, the parameters and fuzzy members are considered variably, which ultimately has turned into an optimization issue. Therefore, the proper coordination between these variables leads to the best operating conditions and vice versa: inappropriate adjustment of the parameters may lead to a persistent exacerbation in applying controlling signals. Because of the complexity of the system and to avoid a large amount of computing, in this paper, optimization of control parameters has been proposed with the help of an improved virus colony search algorithm. The proposed controller has been studied under different operating conditions considering the interregional fluctuation and low frequencies. It is shown that the proposed fuzzy controller has a better performance for stabilizing damaging system disturbances in bad operating conditions.

Collective Intelligence Algorithms in Pedagogical Practice

The article considers the ideas of using artificial intelligence algorithms in pedagogics. It presents the methodology of the so-called collective pedagogical megasystem. The introduction of such an ephemeral construct is necessary only to understand the collective pedagogical intelligence system, formulate it in a model, find out its operation patterns and the laws it obeys. It would contribute to predicting pedagogical processes and phenomena and formulating new laws. The objective of the article is to demonstrate the application of collective intelligence algorithms in pedagogical practice for effective didactic decision-making. The matter is that in a real educational process, besides the well-known set of pedagogical conditions, there are some random and unpredictable reasons and factors that are hard to foresee or anticipate. Due to their stochastic nature, they occur spontaneously. These single reasons make a minor impact on the teaching methods selection, but in aggregate their influence gets so strong that they can upturn some prognostic conclusions. The problem also focuses on identifying the factors that would ensure the highest efficiency and productivity of studies among the known (expected) and random (unexpected) reasons. For these purposes, the most suitable algorithm for the selection training methods is the so-called ant algorithm which, on the one hand, considers the randomness of the influence parameters, and on the other, ensures steady and high productivity. A certain example was selected to demonstrate the process of applying the ant algorithm to reveal the best hierarchy of the pedagogical conditions (factors) that determines the optimum choice of the training method. The authors conclude that human intelligence is distributed and integrated at the same time, and the application of collective intelligence algorithms in pedagogical practice can yield some effective didactic decisions

Investigating the dimensions of globalization and its impact on poverty in Iran: An improved bat algorithm approach

Collective intelligence is one of the most powerful optimization techniques based on group behaviors of organisms. Bat algorithm (BA) is an algorithm inspired by the collective action of bats in the wild, presented in 2010 by Yang. Researchers have made several efforts to improve these algorithms. This article investigates the effect of globalization on Iran's poverty by enhancing the performance of BA. As an inescapable reality, globalization has various political, social, and economic dimensions, each with different effects on poverty. In this article, to improve the algorithm's performance, the speed and motion relationships of bats were modified such that to adapt the movement of bats as optimization solutions toward the target. The mutation operator is also used to check all points of the search space to get rid of the optimal local optimization. The study period is the years 1995 to 2017. The results showed that globalization affects Iran's poverty in various dimensions, and the performance of the improved bat collective intelligence algorithm (ISABA) for modeling is better than that of the bat algorithms (BAs).•This article provides a suitable method for researchers to study poverty.•Improved BAs can help researchers solve complex problems.•The results obtained by the collective intelligence algorithms in this paper help researchers in the field of poverty to compare the results of their research with it.

Comparative Study of Recent Swarm Algorithms for Continuous Optimization

Abstract Optimization is a problem of finding the best solution from various possible solutions. In solving optimization problems, collective intelligence algorithms are often used as a method for finding solutions to optimization problems. This is because collective intelligence has various algorithms to search for various types of continuous optimization problems. However, not all of these algorithms work with the same performance. An algorithm can work better for a problem than other algorithms. Therefore, the diversity of performance of this algorithm must be identified, analyzed, and compared. By knowing the strengths, weaknesses, nature, and behaviour of the algorithm in solving various problems, this is believed to be able to realize the use of more effective algorithms in solving various problems. In this paper, Dragonfly Algorithm (DA), Grey Wolf Algorithm (GWO), and Rao Algorithms are carefully investigated using nine benchmark functions. The result indicates that Rao generally performed better than GWO and DA. DA is outperformed by GWO and Rao in reaching the convergence score, although DA has an edge in searching a large search space, and in theory, given enough population, DA can perform better than GWO or Rao.

Performance analysis of optimized controllers with bio-inspired algorithms

This article proposes tuning of PID controllers through optimization techniques known as collective intelligence algorithms: the colony of ants and cuckoo search, applied to the temperature control of a PCT-2 plant. The dynamic behavior analysis of the controllers tuned by the application of collective intelligence algorithms were performed in comparison to a controller tuned by the genetic algorithm method and a controller designed with the PID tuner tool, obtaining favorable results in the tuning of the algorithms cuckoo search and ants colony in comparison to the PID tuner method and with a faster execution speed than the genetic algorithms method.

Energy-aware virtual machines allocation by krill herd algorithm in cloud data centers

The growing demand for computational power has led to the emergence of large-scale data centers that consume massive amounts of energy, thus resulting in high operating costs and CO2 emission. Furthermore, cloud computing environments are required to provide a high Quality of Service (QoS) to their clients and, therefore, need to handle power shortages. An optimized virtual machine allocation to physical hosts lowers energy consumption and allows for high-quality services. In this study, a novel solution was proposed for the allocation of virtual machines to physical hosts in cloud data centers using the Krill Herd algorithm, which is the fastest collective intelligence algorithm recently introduced. The performance of the proposed method was evaluated using the CloudSim simulator, and the results are suggestive of a 35% reduction in energy consumption.

Social media for e-learning of citizens in smart city

The rapid development of social media can be applied for citizens’ e-learning in a smart city. Big cities have to cope with several open issues like a growing population or a traffic congestion. Especially, a home and public space is supposed to be used in more efficient way. Sustainable homes and buildings can be planned with using some modern techniques. Even currently, there is a huge problem with a lack of key resources like water and energy. So, an efficient management of resources is strongly required. What is more, a global warming and carbon emissions are considered as some critical factors for living conditions in many cities, too. A vision of a smart city is related to a better protection of a natural environment and a more efficient use of it. Moreover, citizens expect an efficient and sustainable transportation in livable city. To present some solutions on above issues, this paper outlines the methodology of using social media to provide necessary knowledge by citizenship training systems. In particular, some selected e-learning applications have been characterized. They are related to pedagogical agents in e-learning. In addition, some advanced meta-heuristics have been proposed with particular emphasis on genetic programming, artificial neural networks, neuro-evolution algorithms, support vector machines, and some collective intelligence algorithms. Finally, cloud services are discussed regarding a smart city management and training.

Implementation of ACO-Based Selection with Backward-Ant Mechanism for Adaptive Routing in Network-on-Chip Systems

The Networks-on-Chip (NoC) provides regular and scalable design architecture for the chip multiprocessor (CMP) systems. The routing efficiency dominates the overall system performance because of more complex applications and network scaling. The Ant Colony Optimization (ACO) is a distributed collective-intelligence algorithm. The ACO-based selection scheme with Backward-Ant mechanism (ACO-BANT) can provide extra feedback congestion information compared with forward-ant mechanism. However, the storing and computation cost of BANT is too high for the NoC systems. In this work, we implement the ACO-BANT selection scheme with feasible cost on NoC. The simulation results show that the proposed scheme yields improvements in saturation throughput by 16.26% compared to the OBL selection. We also implement the router architecture of the proposed scheme, which has the highest improvement-to-overhead ratio.