GA-Based Equipment Support Resource Leveling Optimization

Equipment support decision-making is the important foundation to effectively organize equipment support. Equipment support is dependent on various support resources. The less the fluctuation of resource requirement during the project, the more advantageous the organization and carrying out of equipment support tasks in line with the predetermined schedule. Resource leveling optimization is the basic approach to achieve resource requirement leveling. Traditional exact and heuristic algorithms are inefficient in solving the problems even of medium scale. Genetic algorithm with repairing operator is utilized to overcome drawbacks of traditional methods. The performance of the adopted method is demonstrated by obtaining the optimal or near-optimal solutions to the simulated problems. KEYWORD: Resource leveling; Equipment support; Genetic algorithm; Network planning; Optimization International Conference on Social Science, Education Management and Sports Education (SSEMSE 2015) © 2015. The authors Published by Atlantis Press 805 The resource leveling problem of a project can then be described as follows: given that the duration time of activity (i, j), D(i, j) ( (i, j)W), and the resource requirement per unit time, ) , ( 0 j i r (  (i, j)W); the total project duration, T; the resource requirement quantity of the project at moment t, R(t); the earliest start time, the latest start time, and the actual start time of activity (i, j), tES(i, j), tLS(i, j), and tAS(i, j), respectively. During the process of the optimization, the logical relation among activities should not be changed, an activity not be interrupted once started, and the total project duration be fixed. The objective for resource leveling is to find the actual start time of each activity in the project such that the variance of the resource requirements over the project period is minimized. The model is formulated as an optimization problem as in formula (1a) subject to conditions in formula (1b). Minimize 2 1 2 2 ) ( 1 m T i R t R T      (1a)