Abstract: Mathematics and Information Science, Shaanxi Normal Univ., Xi′an 710062， China)
A novel trust model based on the L-fuzzy set is established. It describes the trust between entities by linguistic variables, presents a formalized depiction based on context-free grammar and uses fuzzy operators to dispose trust levels. A new task scheduling algorithm L-FSTM-Min-Min (L-Fuzzy Set Trust Model-Min-Min) is brought out based on this model and Min-Min algorithm. This algorithm utilizes the triple I method to make fuzzy inference for multi-recommendation trust rules based on fuzzy logic and achieves comprehensive trust evaluation.Through the experiment by the GridSim simulator, results show that task completing time is obviously reduced and that the number of failed services is greatly decreased compared with the Min-Min algorithm in the same condition.

Key words: fuzzy set theory, linguistic variables, trust model, scheduling algorithm