Energy–time modelling of distributed multi-population genetic algorithms with dynamic workload in HPC clusters

Resource optimization with MPI process malleability for dynamic workloads in HPC clusters

Cloud platforms often rely on reactive, threshold-based auto-scaling, which can lead to both over-provisioning (wasted cost) and under-provisioning (performance degradation) under dynamic workloads. We present a fully integrated framework that forecasts short-term resource demands using hybrid time-series models (LSTM neural networks + ARIMA) and drives proactive scaling decisions via a dual-stage optimizer combining Deep Q-Learning (DQN) and Genetic Algorithms (GA). Deployed on a local Kubernetes testbed, our solution achieves over 90 % forecasting accuracy (RMSE < 0.05), reduces operational cost by ~25 %, and improves average CPU utilization from 60 % to 85 %, while maintaining sub-200 ms scaling latencies. This hybrid approach also yields an estimated 15%energy savings by minimizing idle resources—demonstrating a practical path toward cost- and energy-efficient cloud resource management.

Cloud-based software services necessitate adaptive resource allocation with the promise of dynamic resource adjustment for guaranteeing the Quality-of-Service (QoS) and reducing resource costs. However, it is challenging to achieve adaptive resource allocation for software services in complex cloud environments with dynamic workloads. To address this essential problem, we propose an adaptive resource allocation strategy for cloud-based software services with workload-time windows. Based on the QoS prediction, the proposed strategy first brings the current and future workloads into the process of calculating resource allocation plans. Next, the particle swarm optimization and genetic algorithm (PSO-GA) is proposed to make runtime decisions for exploring the objective resource allocation plan. Using the RUBiS benchmark, the extensive simulation experiments are conducted to validate the effectiveness of the proposed strategy on improving the performance of resource allocation for cloud-based software services. The simulation results show that the proposed strategy can obtain a better trade-off between the QoS and resource costs than two classic resource allocation methods.

Embedded systems continue to execute computational- and memory-intensive applications with vast data sets, dynamic workloads, and dynamic execution characteristics. Adaptive distributed and heterogeneous embedded systems are increasingly critical in supporting dynamic execution requirements. With pervasive network access within these systems, security is a critical design concern that must be considered and optimized within such dynamically adaptive systems. This paper presents a modeling and optimization framework for distributed, heterogeneous embedded systems. A dataflow-based modeling framework for adaptive streaming applications integrates models for computational latency, mixed cryptographic implementations for inter-task and intra-task communication, security levels, communication latency, and power consumption. For the security model, we present a level-based modeling of cryptographic algorithms using mixed cryptographic implementations. This level-based security model enables the development of an efficient, multi-objective genetic optimization algorithm to optimize security and energy consumption subject to current application requirements and security policy constraints. The presented methodology is evaluated using a video-based object detection and tracking application and several synthetic benchmarks representing various application types and dynamic execution characteristics. Experimental results demonstrate the benefits of a mixed cryptographic algorithm security model compared to using a single, fixed cryptographic algorithm. Results also highlight how security policy constraints can yield increased security strength and cryptographic diversity for the same energy constraint.

In this investigation a step-wise “cross-evaluation” procedure has been implemented aiming to assess the quality of multi-population genetic algorithms (MpGA) performance. Three MpGA, searching for an optimal solution applying main genetic operators selection, crossover and mutation in different order, have been here applied in such a challenging object as parameter identification of a fermentation process model. The performance quality of standard MpGA algorithm, denoted as MpGA_SCM (coming from selection, crossover, mutation), and two modifications, respectively MpGA_MCS (mutation, crossover, selection) and MpGA_CMS (crossover, mutation, selection) have been investigated for the purposes of parameter identification of S. cerevisiae fed-batch cultivation. As an alternative to conventional criteria for assessing the quality of algorithms performance, here an intuitionistic fuzzy logic (IFL) is going to be implemented. Also, this is the first time when two modifications of standard MpGA_SCM, in which the selection operator is performed as the last one, after crossover and mutation, are going to be evaluated. The performance of three MpGA is going to be assessed applying a step-wise procedure implementing IFL. As a result, MpGA_SCM has been approved as a leader between three considered here MpGA. The leadership between MpGA_CMS and MpGA_MCS depends on the researcher choice between a bit slower, but more highly evaluated MpGA_CMS towards faster one, but a bit less highly evaluated MpGA_MCS.

This study develops an energy management strategy (EMS) for hybrid combine harvesters to address fluctuating power demands in agricultural operations. By segmenting harvesting processes into quasi-periodic cycles linked to machine dynamics, the method integrates component-specific power models (header, conveyor, drum) for accurate energy estimation. Real-time feed rate adjustments are achieved through dynamic responses of critical components, optimizing cycle duration and power allocation. A genetic algorithm synchronizes energy distribution and cycle timing to minimize fuel consumption. Validated via AMESim/Simulink co-simulation with dual engine models, the strategy reduces fuel use by 21.1% compared to conventional systems. Key innovations include quasi-periodic load segmentation, component-response-based feed rate prediction, and GA-driven multi-objective optimization. The approach enhances adaptability to variable harvesting conditions, offering a scalable framework for energy-efficient electrification in agriculture. Results demonstrate significant potential for hybrid systems in reducing operational costs and emissions while maintaining productivity under dynamic workloads.

Multi-population genetic algorithms (MGAs), extensions of traditional single-population genetic algorithms (SGAs), have been recognized as being more effective both in speed and solution quality than SGAs. Despite of these advantages, the behavior and performance of MGAs, like SGAs, are still heavily affected by an appropriate choice of parameters such as connection topology, migration method, population number, migration interval, etc. In the past few years, though some researchers have investigated schemes for automating the parameter settings for MGAs, no work, to our knowledge, has ever investigated self-adaptation in MGAs in a systematic way. In this paper, we survey the previous work and categorize the self-adaptation of MGAs into three aspects. According to this classification, we introduce a systematic research roadmap for investigating the self-adaptation of MGAs.

In recent years, multi-population genetic algorithms (MGAs) have been recognized as being more effective both in speed and solution quality than single-population genetic algorithms (SGAs). Despite of these advantages, the behavior and performance of MGAs, like SGAs, are still heavily affected by a judicious choice of parameters, such as connection topology, migration method, migration interval, migration rate, population number, etc. In this paper, the issue of adapting migration parameters for MGAs is investigated. We examine, in particular, the effect of adapting the migration interval as well as migration rate on the performance and solution quality of MGAs. Thereby, we propose an adaptive scheme to evolve the appropriate migration interval and migration rate for MGAs. Experiments on the 0/1 knapsack problem showed that our approach can compete with MGAs with static migration parameters.

The high popularity and growing demand of cloud computing has a strong effect on the cloud infrastructure providers to efficiently man-age their cloud datacenters in order to fulfill provisioning of everything in the form of a service to end users and also to achieve efficient balancing between its less energy consumption for reduced environmental affects and maximize revue. This paper presents an energy efficient framework for green cloud datacenter which considers resource utilization and energy efficiency to support resource allocation decisions towards green computing. This work mainly relies on energy efficient provisioning of resources utilizing an application predic-tion and VM provisioning mechanism using genetic algorithm. Our approach has been validated by performing a set of experiments un-der dynamic cloud environment workload scenarios using Cloudsim toolkit. Compared to the benchmark (existing) algorithms, our method is able to significantly reduce the energy consumption cost without a priori knowledge of the future workloads

Intuitionistic fuzzy logic is the main tool in the recently developed step-wise “cross-evaluation” procedure that aims at the assessment of different optimization algorithms. In this investigation, the procedure previously applied to compare the effectiveness of two or three algorithms has been significantly upgraded to evaluate the performance of a set of four algorithms. For the first time, the procedure applied here has been tested in the evaluation of the effectiveness of genetic algorithms (GAs), which are proven as very promising and successful optimization techniques for solving hard non-linear optimization tasks. As a case study exemplified with the parameter identification of a S. cerevisiae fed-batch fermentation process model, the cross-evaluation procedure has been executed to compare four different types of GAs, and more specifically, multi-population genetic algorithms (MGAs), which differ in the order of application of the three genetic operators: Selection, crossover and mutation. The results obtained from the implementation of the upgraded intuitionistic fuzzy logic-based procedure for MGA performance assessment have been analyzed, and the standard MGA has been outlined as the fastest and most reliable one among the four investigated algorithms.

Modern embedded systems are becoming increasingly multifunctional, and, as a consequence, they more and more have to deal with dynamic application workloads. This dynamism manifests itself in the presence of multiple applications that can simultaneously execute and contend for resources in a single embedded system as well as the dynamic behavior within applications themselves. Such dynamic behavior in application workloads must be taken into account during the early system-level Design Space Exploration (DSE) of Multiprocessor System-on-Chip (MPSoC)-based embedded systems. Scenario-based DSE utilizes the concept of application scenarios to search for optimal mappings of a multi-application workload onto an MPSoC. To this end, scenario-based DSE uses a multi-objective genetic algorithm (GA) to identify the mapping with the best average quality for all the application scenarios in the workload. In order to keep the exploration of the scenario-based DSE efficient, fitness prediction is used to obtain the quality of a mapping. This fitness prediction implies that instead of using the entire set of all possible application scenarios, a small but representative subset of application scenarios is used to determine the fitness of mapping solutions. Since the representativeness of such a subset is dependent on the application mappings being explored, these representative subsets of application scenarios are dynamically obtained by means of coexploration of the scenario subset space. In this chapter, we provide an overview of scenario-based DSE and, in particular, present multiple techniques for fitness prediction using representative subsets of application scenarios: a stochastic, deterministic, and hybrid combination.

Multi-population genetic algorithms have been used with success for several multi-objective optimization problems. In this paper, we present a new general multi-population genetic algorithm for evolving decision trees. It was designed to improve the possibility of evolving balanced decision trees, simultaneously optimized for the predictions of each class. Single-population genetic algorithms namely tend to construct decision trees with great variance in single class accuracies. The proposed approach is tested over 10 UCI datasets, and it is compared with a single-population genetic algorithm as well as with traditional decision-tree induction algorithms. Results show that the designed multi-population approach provides classification results comparable to C4.5 and CART in terms of accuracy and tree size, while outperforming them regarding balanced solutions (in terms of average class accuracy and range of single-class accuracies).

Among the several variants of GA, multi-population genetic algorithms (MGA) have created a niche in the literature of genetic algorithms owing to their ability to explore the global optima. But the problem associated with them is the judicious choice of the migration parameters, which when not chosen properly may lead to performance degradation of MGA. This paper discusses the advantages of an adaptive multi-population GA called nomadic genetic algorithm (NGA) with respect to its migration policies and highlights its betterment over other algorithms of the type.

A variety of previous works exist on maintaining population diversity of genetic algorithms (GAs). Dual-population GA (DPGA) is a type of multipopulation GA (MPGA) that uses an additional population as a reservoir of diversity. The main population is similar to that of an ordinary GA and evolves to find good solutions. The reserve population evolves to maintain and provide diversity to the main population. While most MPGAs use migration as a means of information exchange between different populations, DPGA uses crossbreeding because the two populations have entirely different fitness functions. The reserve population cannot provide useful diversity to the main population unless the two maintain an appropriate distance. Therefore, DPGA adjusts the distance dynamically to achieve an appropriate balance between exploration and exploitation. The experimental results on various classes of problems using binary, real-valued, and order-based representations show that DPGA quite often outperforms not only the standard GAs but also other GAs having additional mechanisms of diversity preservation.

Populations of multipopulation genetic algorithms (MPGAs) parallely evolve with some interaction mechanisms. Previous studies have shown that the interaction structures can impact on the performance of MPGAs to some extent. This paper introduces the concept of complex networks such as ring-shaped networks and small-world networks to study how interaction structures and their parameters influence the MPGAs, where subpopulations are regarded as nodes and their interaction or migration of elites between subpopulations as edges. After solving the flexible job-shop scheduling problem (FJSP) by MPGAs with different parameters of interaction structures, simulation results were measured by criteria, such as success rate and average optimal value. The analysis reveals that (1) the smaller the average path length (APL) of the network is, the higher the propagation rate will be; (2) the performance of MPGAs increased first and then decreased along with the decrease of APL, indicating that, for better performance, the networks should have a proper APL, which can be adjusted by changing the structural parameters of networks; and (3) because the edge number of small-world networks remains unchanged with different rewiring possibilities of edges, the change in performance indicates that the MPGA can be improved by a more proper interaction structure of subpopulations as other conditions remain unchanged.