[Naming Convention] CourseCode-StartDate-BatchCode-Group_or_Individual-TeamName_or_PersonName-ProjectName.zip
-
[MTech Group Project Naming Example] IRS-MR-2019-01-19-IS1PT-GRP-AwsomeSG-HDB_BTO_Recommender.zip
-
[MTech Individual Project Naming Example] IRS-MR-2019-07-01-IS1FT-IND-SamGuZhan-HDB_BTO_Process.zip
-
[EEP Group Project Naming Example] IRS-MR-2019-03-13-EEP-GRP-AwsomeSG-HDB_BTO_Recommender.zip
-
[EEP Individual Project Naming Example] IRS-MR-2019-08-22-EEP-IND-SamGuZhan-HDB_BTO_Process.zip
Online editor for this README.md markdown file
Abstract
We developed an IT project release schedule optimizer using Genetic Algorithm, considering various real world constraints, including human resources, lead time and software version compatibility. The goal for optimizer is to shorten the overall projects' life span as well as individual project's life spans for a group of concurrently running projects. The implementation tools used are Microsoft Excel and Evolver.
Keywords: Genetic Algorithm, Project Release Management, Optimization, Evolver
Problem Introduction
Release Management is the process of managing software (IT project) releases from development stage to software release in production. Software products (with different versions/upgrades) are typically in an ongoing cycle of planning, development, various testing, and production release.
There is growing complexity of dynamic project dependencies, maximum available resources, and large number of parallel running projects, together with various moving/changing influencing factors to take into consideration. These factors must fit together seamlessly to guarantee success and long-term value of multiple running projects as a whole.
We developed a simplified GA optimizer to conduct multiple project releases scheduling (to shorten make-span of both individual project and all projects as a whole), considering several resource constraints, including available human resources, testing environments, and software version conflicts among all projects.
| Official Full Name | Student ID (MTech Applicable) | Work Items (Who Did What) | Email (Optional) |
|---|---|---|---|
| GU Zhan (Sam) | A0107682A | domain modeling, fitness function design, score integraion | [email protected] |
| LIU Jie (Judy) | A0107579U | execuate GA runs, draft report | [email protected] |
| Francisco Liwa | A0107538A | execuate GA runs, version control | [email protected] |
Note: It is not mandatory for every project member to appear in video presentation; Presentation by one project member is acceptable. More reference video presentations here
<Github File Link> : https://github.com/telescopeuser/Workshop-Project-Submission-Template-GA-Optimizer/blob/master/UserGuide/Refer%20to%20Project%20Release%20Optimizer.xlsx.txt
This project was implemented using a properiatory software: Evolver, which is a plug-in tool for Microsoft Excel.
[ 2014 September ] The Excel-Evolver implemetation of Project Release Optimizer can be found in \SystemCode\Excel Evolver\Project Release Optimizer.xlsx
The free iss-vm workstation also ships a free-of-charge evolutionary nonlinear programming solver: NLP Solver, as a plug-in for LibreOffice Calc (Microsoft Excel alike).
The NLP Solver contains below algorthms/models relating to evolutionary programming:
1. DEPS - Differential Evolution & Particle Swarm Optimization
DEPS consists of two independent algorithms: Differential Evolution and Particle Swarm Optimization. Both are especially suited for numerical problems, such as nonlinear optimization, and are complementary to each other in that they even out their others shortcomings.
The idea behind Particle Swarm Optimization is to represent every solution vector as coordinates in an n-dimensional room. Each individual (particle) traverses through that space and keeps track of its own best point so far. This information as well as the knowledge about the particle with the best current solution influence how it will move in each iteration. Due to that, the particles will always try to build a swarm and float around the best solutions, traversing possible even better solutions on its way.
Differential Evolution on the other hand is a strategy to "recombine" two individuals similar to Genetic Algorithms. Instead of crossing over chromosomes (i.e. mixing up their variables), the knowledge about the target function is used to anneal both points.
In each iteration of the algorithm, each individual chooses one of both strategies and applies them to its current solution vector. The probability, which strategy is chosen, can be modified with the option Agent Switch Rate.
2. SCO - Social Cognitive Optimization
SCO takes into account the human behavior of learning and sharing informations. Each individual has access to a common library with knowledge shared between all individuals.
In each step, an individual looks up the (presumably) best information available in the library and builds a decision based on it together with its own current knowledge about the problem. Afterwards it replaces one of the worse informations from the library with its newly learned solution.
That way all individuals work together (the social aspect) and make up their own mind (the cognitive aspect). Therefore it's called Social Cognitive Optimization.
[ 2019 March ] The NLP-Solver implemetation of Project Release Optimizer can be found in \SystemCode\ISS-VM Solver\Project Release Optimizer iss-vm.ods
<Github File Link> : https://github.com/telescopeuser/Workshop-Project-Submission-Template-GA-Optimizer/blob/master/ProjectReport/GA%20Optimizer%20Report.pdf
Recommended Sections for Project Report / Paper:
- Executive Summary / Paper Abstract
- Sponsor Company Introduction (if applicable)
- Business Problem Background
- Project Objectives & Success Measurements
- Project Solution (To detail domain modelling & system design.)
- Project Implementation (To detail system development & testing approach.)
- Project Performance & Validation (To prove project objectives are met.)
- Project Conclusions: Findings & Recommendation
- List of Abbreviations (if applicable)
- References (if applicable)
- Visual summary of below Evaluation Statistics
- Raw evaluation results for respective project release optimization GA runs, shown below:
We evaluate the GA model in two main approaches, with same initial fitness (project data inputs) for all GA runs:
-
Conduct Fixed, Mutation-Adaptive and Crossover-Adaptive GA runs for ten rounds respectively, at 5 minutes per GA run. Thus in total there are 30 GA runs, covering 150 minutes or 2.5 hours running time.
-
Conduct one round of GA run with respective longer time limits: 10, 20, 30, 40, 50, 60 minutes, and 2, 4, 6, 8 hours. Thus in total there are 10 GA runs, covering 1,360 minutes or 22.67 hours running time.
This NICF - Reasoning Systems (SF) course is part of the Analytics and Intelligent Systems and Graduate Certificate in Intelligent Reasoning Systems (IRS) series offered by NUS-ISS.
Course Manager: GU Zhan (Sam)
