Efficiently Solving: A Large-Scale Integer Linear Program Using a Customized Genetic Algorithm

Deb, Kalyanmoy; Pal, Koushik

doi:10.1007/978-3-540-24854-5_104

“…In this approach, release planning is formalized as an integer linear programming problem; however, because it is difficult and too expensive to explore all feasible solutions using traditional linear programming techniques [42], genetic algorithm (GA) is used to find the optimal or near-optimal solutions. GA [43] proposing EVOLVE*, which consists of three main phases: 1) modeling, where the problem is formalized as an ILP problem 2) exploration, where a genetic algorithm is used to produce a set of potential solutions and 3) consolidation, where the solutions produced in Stage 2 are evaluated by the release management.…”

Section: Combination Of Linear Programming and Genetic Algorithmmentioning

confidence: 99%

“…Each chromosome consists of gens and each gen contains a trait. Genetic Algorithm (GA) is population-based search algorithm that search for the optimal or near optimal solutions by producing a sequence of generations [42].…”

Section: Genetic Algorithm-based Approachmentioning

confidence: 99%

Release planning for multi-tenant software as a service (SaaS) applications

Alrashoud¹

2021

Preprint

0

View full text Add to dashboard Cite

In multi-tenant Software as a Service (SaaS) applications, the providers are required to regularly deliver new releases of the software in order to satisfy the evolving requirements of tenants. The first step in a release development lifecycle is the release planning process. This thesis formulates the problem of the "next release" planning for multi-tenant Software as a Service (SaaS) applications. Two variables that influence release planning in SaaS applications are introduced: the degree of commonality of features and the contractual constraints. The commonality of a feature denotes the number of tenants that have requested that feature. The contractual constraints denote the effects of service levels to which tenants have subscribed on the release planning process. Furthermore, this thesis proposes three novel approaches in order to tackle the problem of the "next release" planning for multi-tenant SaaS applications. The first one is a prioritization approach that employs a Fuzzy Inference System (FIS) engine in order to speed up the release planning process and overcome the uncertainty associated with the human judgment. In this approach, the human expertise, which is represented by fuzzy rules, is considered automatically in the release planning process. The second and third approaches consider release planning as an optimization problem. The second approach uses an exact optimization method (Binary Linear Programming (BLP)) in order to generate an optimal release plan, while the third approach uses heuristic optimization method (Genetic Algorithm (GA)). All of the three approaches aim to generate a plan for the next release that maximizes the degree of overall tenants’ satisfaction, maximizes the degree of commonality, and minimizes the potential risk while taking into account contractual, effort, and dependencies constraints. Moreover, the thesis presents an experimental study of the proposed approaches in order to determine which approach is best suited to different sets of scenarios. In this experiment, the performance of the proposed approaches is evaluated using four criteria: the overall tenants’ satisfaction, the commonality, the adherence to the risk, and the running time. Additionally, the thesis presents an experiment that compares the proposed approaches with a compared model that is selected from the literature.

show abstract

“…In this approach, release planning is formalized as an integer linear programming problem; however, because it is difficult and too expensive to explore all feasible solutions using traditional linear programming techniques [42], genetic algorithm (GA) is used to find the optimal or near-optimal solutions. GA [43] proposing EVOLVE*, which consists of three main phases: 1) modeling, where the problem is formalized as an ILP problem 2) exploration, where a genetic algorithm is used to produce a set of potential solutions and 3) consolidation, where the solutions produced in Stage 2 are evaluated by the release management.…”

Section: Combination Of Linear Programming and Genetic Algorithmmentioning

confidence: 99%

Release planning for multi-tenant software as a service (SaaS) applications

Alrashoud¹

2021

Preprint

0

View full text Add to dashboard Cite

In multi-tenant Software as a Service (SaaS) applications, the providers are required to regularly deliver new releases of the software in order to satisfy the evolving requirements of tenants. The first step in a release development lifecycle is the release planning process. This thesis formulates the problem of the "next release" planning for multi-tenant Software as a Service (SaaS) applications. Two variables that influence release planning in SaaS applications are introduced: the degree of commonality of features and the contractual constraints. The commonality of a feature denotes the number of tenants that have requested that feature. The contractual constraints denote the effects of service levels to which tenants have subscribed on the release planning process. Furthermore, this thesis proposes three novel approaches in order to tackle the problem of the "next release" planning for multi-tenant SaaS applications. The first one is a prioritization approach that employs a Fuzzy Inference System (FIS) engine in order to speed up the release planning process and overcome the uncertainty associated with the human judgment. In this approach, the human expertise, which is represented by fuzzy rules, is considered automatically in the release planning process. The second and third approaches consider release planning as an optimization problem. The second approach uses an exact optimization method (Binary Linear Programming (BLP)) in order to generate an optimal release plan, while the third approach uses heuristic optimization method (Genetic Algorithm (GA)). All of the three approaches aim to generate a plan for the next release that maximizes the degree of overall tenants’ satisfaction, maximizes the degree of commonality, and minimizes the potential risk while taking into account contractual, effort, and dependencies constraints. Moreover, the thesis presents an experimental study of the proposed approaches in order to determine which approach is best suited to different sets of scenarios. In this experiment, the performance of the proposed approaches is evaluated using four criteria: the overall tenants’ satisfaction, the commonality, the adherence to the risk, and the running time. Additionally, the thesis presents an experiment that compares the proposed approaches with a compared model that is selected from the literature.

show abstract

“…Over the recent years, there has been a push to develop optimization procedures to tackle large-scale problems [2,3,5,4,10,14,19,20,25,24]. Many of the aforementioned large-scale optimization studies have relied on linear programming solvers such as simplex methods and interior point methods.…”

Section: Literature Reviewmentioning

confidence: 99%

“…While these methods are very efficient in solving linear programming problems, they fall quite short of solving nonlinear, noisy, and deceptive problems. To the best of our knowledge, the largest known GA in the literature solves a problem with 4-million binary variables [5,4,24], but its applicability to other optimization problems remains in doubt because of problem-specific operators, small population used, and the lack of theory.…”

Section: Literature Reviewmentioning

confidence: 99%

Towards billion-bit optimization via a parallel estimation of distribution algorithm

Sastry

¹

,

Goldberg

²

,

Llorà

³

2007

Proceedings of the 9th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

This paper presents a highly efficient, fully parallelized implementation of the compact genetic algorithm (cGA) to solve very large scale problems with millions to billions of variables. The paper presents principled results demonstrating the scalable solution of a difficult test function on instances over a billion variables using a parallel implementation of cGA. The problem addressed is a noisy, blind problem over a vector of binary decision variables. Noise is added equaling up to a tenth of the deterministic objective function variance of the problem, thereby making it difficult for simple hillclimbers to find the optimal solution. The compact GA, on the other hand, is able to find the optimum in the presence of noise quickly, reliably, and accurately, and the solution scalability follows known convergence theories. These results on noisy problem together with other results on problems involving varying modularity, hierarchy, and overlap foreshadow routine solution of billion-variable problems across the landscape of search problems.

show abstract

Efficiently Solving: A Large-Scale Integer Linear Program Using a Customized Genetic Algorithm

Cited by 7 publications

References 5 publications

Release planning for multi-tenant software as a service (SaaS) applications

Release planning for multi-tenant software as a service (SaaS) applications

Release planning for multi-tenant software as a service (SaaS) applications

Towards billion-bit optimization via a parallel estimation of distribution algorithm

Contact Info

Product

Resources

About