The problem of jointly designing frequencies and routes is also widely treated. Van Nes et al. (1988) proposed a method to
jointly design routes and frequencies with the aim of serving demand and maximising direct trips; a heuristic algorithm was
proposed for solving the problem, starting from all possible routes with null frequency and increasing the frequencies
according to their marginal efficiency. Pattnaik et al. (1998) proposed a genetic algorithm for solving the transit routes
and frequency optimisation problem; the objective was to minimise both operator costs and users travel times and the genetic
algorithm was adopted for generating transit routes. Carrese and Gori (2002) described a complex urban bus network
design procedure and proposed a sequential heuristic procedure to design routes and frequencies so as to minimise the sum
of users and operator costs. Ceder (2003) proposed a model and a heuristic demand-oriented approach for designing routes
and frequencies in two stages: in the first stage all routes and transfers are generated while in the second a short-turning
strategy is applied to reduce the length of some runs, in order to lower operator costs. Tom and Mohan (2003) proposed
a genetic algorithm that simultaneously designs routes and frequencies minimising both operator and users costs; in this
case the frequency of a line is included as a variable in the coding scheme.