Resource Leveling With Fixed Project Duration In Software Industry Based On Ant Colony Algorithm And Optimization Decision

The software industry as a knowledge intensive and resource intensive industry, project management is more difficult compared with general industry. Resource leveling has important practical significance to improve the utilization rate of resources, control the construction cost, improve the quality of project construction and ensure the project schedule plan. In addition, from the point of the research status at home and abroad, the use of intelligent bionic algorithm for solving the resource leveling problem mainly confined to the application of genetic algorithm, instead of using ant colony algorithm on the research. This paper designs the ant colony algorithm solving process, verifies the validity of the ant colony algorithm in solving resource leveling, and it has important academic significance.This article is based on the fixed project duration- resource leveling problem, and the research status at home and abroad and the theoretical analysis. First, this paper introduces the theoretical basis of project management, network planning and resource. Secondly, this paper introduces the fixed project duration- resource leveling common evaluation index, including variance, deviation, maximum absolute deviation and uneven coefficient, and on this basis to build the mathematical model of this paper: using the standard deviation of the weighted as objective optimization function, and the starting time, relaxation time and operation sequence three aspects as constraint conditions. Then, this paper introduces the basic theory of ant colony algorithm, and put forward the ant colony algorithm of solving the problem of fixed project duration-resource leveling based on positive duration of the basic ant colony algorithm. Solve the non-critical process work start time scope; all the ants randomly distributed in the feasible region above, get the initial position of ants, and calculate the value of their corresponding information, and get the optimal value. Ants according to the number of the transition probability to local search or global search, until the number of iterations reaches maximum cycle times, and obtain the global optimal solution. At last, take empirical test to prove the effectiveness of the algorithm. Take a single resource leveling and multi-resource leveling for examples, and shows the algorithm design and the results of the two examples analysis.The study found that after ant colony algorithm to solve the calculation, the objective function of the single resource value reduced from 10.1225 to 4.3261, the objective function value of more resources from 2.5546 down to 1.7613. Using genetic algorithm to solve the two examples of the paper and compares to the ant colony algorithm, it can be obtained: ant colony algorithm for single resource leveling calculation need 300 species, 1000 iterations, takes 90.473 seconds, using genetic algorithm to solve this case to need 300 species, 1500 iterations, takes 146.470 seconds; Use genetic algorithm of more resource leveling calculation, needs to be 25 species, 100 times iteration, takes 3.860 seconds, using ant colony algorithm, need 25 species, 100 times iteration, takes 1.125 seconds. Compared with the evolution of ant colony algorithm and the genetic algorithm, this paper puts forward that the ant colony algorithm can not only effectively resolve the fixed time limit for a project- resource leveling problem with genetic algorithm greatly shorten the time.Research paper gets some information of the software industry has a certain meaning, is expected to push the software industry resource leveling, achieve greater economic benefits, and expect some help to the further study.