无法在R中使用genalg解决运输优化

我试图使用R中的

lpsolve

和

genalg

解决以下运输问题。我使用

genalg

面临一些问题

            Boston  New York    Supply
---------------------------------------  
Detroit     30      20          200  
Pittsburgh  40      10          100
---------------------------------------  
Demand      150     150 

如果上述问题不清楚，这里是问题陈述的链接。

使用genalg编码：

    library(genalg)  
    dataset <- data.frame(arc = c("det_bos", "det_new", "pit_bos", 
   "pit_new"), cost = c(30, 20, 40, 10), supply_1 = c(1,1,0,0),
    supply_2 = c(0,0,1,1),demand_1 = c(1,0,1,0),demand_2 = c(0,1,0,1))

 supply_1_limit <- 200  
 supply_2_limit <- 100  
 demand_1_limit <- 150  
 demand_2_limit <- 150

evalFunc <- function(arc=c()) {  
current_solution_cost <- arc%*%dataset$cost  
current_solution_supply_1 <- arc%*%dataset$supply_1  
current_solution_supply_2 <- arc%*%dataset$supply_2  
current_solution_demand_1 <- arc%*%dataset$demand_1  
current_solution_demand_2 <- arc%*%dataset$demand_2

if (current_solution_supply_1 < supply_1_limit | 
    current_solution_supply_2 < supply_2_limit | 
    current_solution_demand_1 > demand_1_limit | 
    current_solution_demand_2 > demand_2_limit)    
    return(0) else return(current_solution_cost)}

 GAmodel <- rbga(stringMin=c(0,0,0,0), 
 stringMax=c(150,150,150,150),       
 suggestions=NULL,popSize=200, 
 iters=100, mutationChance=0.10,elitism=T,
 monitorFunc=NULL,evalFunc=evalFunc,showSettings=FALSE, verbose=FALSE)

 cat(summary.rbga(GAmodel))

答案应该是

Total Cost = 6500 , Units Shipped = 300  as ( 150 , 50, 0 , 100)

---

请帮忙

您提到您正在使用lpSolve。你知道lpSolve有一个专门的接口来解决交通问题吗？加载lpSolve并查看

？lp.运输

下面是您的示例问题的实现：

library(lpSolve)
cost.mat  <- matrix(c(30,20,40,10), nrow=2,byrow=TRUE)
row.signs <- c("=","=")
col.signs <- c("=","=")
col.rhs <- c(150,150)
row.rhs <- c(200,100)
sol<-lp.transport (cost.mat, direction="min", row.signs, row.rhs, 
            col.signs, col.rhs, presolve=0,  compute.sens=0, integers = 1:4 )

---

与SAS结果相符。

感谢您的回复。我们已经用lpsolve解决了这个问题。现在，我们正在尝试使用元启发式来解决它，从而尝试genalg的方法。如果可以使用genalg解决此问题，则会很有帮助。

library(lpSolve)
cost.mat  <- matrix(c(30,20,40,10), nrow=2,byrow=TRUE)
row.signs <- c("=","=")
col.signs <- c("=","=")
col.rhs <- c(150,150)
row.rhs <- c(200,100)
sol<-lp.transport (cost.mat, direction="min", row.signs, row.rhs, 
            col.signs, col.rhs, presolve=0,  compute.sens=0, integers = 1:4 )

> sol$solution
     [,1] [,2]
[1,]  150   50
[2,]    0  100