Job-housing relationship in city has always been a concern of scholars. Excess commuting is a well-accepted concept for evaluating commuting efficiency and potential of optimizing it. By calculating difference between observed commute and theoretical minimum commute, excess commuting can measure to what extent the commute of a city is wasted. Studies on excess commuting are rather scarce in China, especially in perspective of different age groups. Using the mobile phone grid data in September 2017 in Shanghai, this paper analyzes the residential locations, job locations, and commuting distances of three age groups, namely, the young, the middle and the old. To calculate their excess commuting, a modified algorithm is devised, simulating individuals exchanging their residences based on the principle of Pareto Optimality, so that no one has to sacrifice their status quo. The new method also uses large-scale individual data instead of small-scale aggregate origin-destination data adopted in the traditional linear programming method. The job-housing relationships under the optimal commuting are estimated and compared between age groups. The results show that: (1) There is an evident inter-generational differentiation in the job-housing relationships. The young people are the disadvantaged group with the longest commute distances and the farthest residence locations from the city center; the elderly are the dominant group with the shortest commute distances and the closest residence locations from the city center; the middle-aged are in-between. (2) The excess commuting rate in Shanghai is 69%, indicating a large potential in optimizing the job-housing relationships. (3) The benefits of the optimization are immense: the average commuting distance declines, the inter-generational differentiation moderates, and the spatial structure of population age becomes more balanced. (4) The young group will benefit most from this process. It is suggested that the spatial planning and policy making of Shanghai should set a target for an ideal spatial structure of the urban population age, so as to increase the fluidity of the job and housing spaces, youthen the city center, and make the perimeter areas more livable for the middle-age and old people.