Genetic algorithm scheduling

Abstract: For some specific types of construction projects, the classical CPM or PDM scheduling techniques are not the most suitable. Few specific scheduling approaches have been developed to cope with construction projects that are made of either repetitive activities or activities with linear developments. But real‐world construction projects do not consist only of such activities. They are generally made of a mixture of linear and/or repetitive activities and of more conventional activities. To allow this, the linear scheduling problem is reformulated, so classical schedule calculation approaches can be used. The implementation of some Allen's algebra features to avoid adverse discontinuities and to allow crew/work continuity, together with a resource‐driven and space‐constrained scheduling are among the key features of the proposed approach. It is also a spin‐off of off‐the‐field practices used for scheduling real projects in the particle accelerator construction domain; an excerpt from such a construction proj...

Abstract: This paper integrates multi-project scheduling and linear scheduling concepts. Since the problem is combinatorial, a two-stage heuristic solution-finding procedure is used to model the problem with multiple resource constraints. Simulated annealing is utilized as a searching engine in the second stage to find the probable optimized solution. The first stage is slightly different from the other two-stage solution finding procedures which are proposed till now. A numerical example of a multi-project situation is given and solved as well.

Abstract: In this paper, the multi-objective flexible job-shop scheduling problem (FJSP) is studied and formulated in a mixed integer programming model. Combining the artificial immune mechanism and genetic algorithm, an immune genetic algorithm is proposed to solve the FJSP. It can generate excellent individuals based on coding and heuristic rules in the initial population and combine the Pareto-optimality and random to deal with the multiple objectives of the FJSP. The immune mechanism is applied to increase the chaotic sequence search feature in order to improve the diversity in process of genetic evolution which can generate better solution. Numerical experiments demonstrate the effectiveness and efficiency of the algorithm.