Current normal transit network results in long passenger travel time because it does not consider the particularity of the travel demand in transit network between urban area and suburb. To solve this problem

a novel transit network model was proposed based on principles of milk-run and hub-spoke

which considers the capacity constraint for the travel demand in transit network between urban area and suburb. Then the corresponding genetic algorithm was developed to solve this model. In this model

the number and locations of the hub stops

the route structure of milk-run route

and the vehicle assignment can be determined. Different from normal transit network

decentralized passenger flow can be aggregated at hub stops to form scale effect through milk-run routes in this network. In addition

massive passenger flow can complete their travels by express routes from hub stops

which can reduce their travel time. Finally

the proposed approach was applied to a real network in Tin Shui Wai of Hong Kong to verify its effectiveness. The comparisons of the proposed network and the existing network show that when the proposed network was applied

the total travel time was reduced by 16.26% with existing available fleet size. It suggests that the proposed network can improve the existing transit service level and travel satisfaction

as well as attract more passengers to travel by public transit

thereby mitigating traffic jam.