Discussion 



The basic optimization program has the capabil- 

 ity and inherent versatility to be utilized for a 

 wide range of applications. The round-trip capa- 

 bility can be modified to a one-way path calcula- 

 tion as was done for the East Delta portion of the 

 groundfish survey by manipulating the distance 

 matrix. Cartesian integrity of the start-stop points 

 is kept intact but the distance between the two 

 stations is set equal to zero in the distance matrix. 

 The program then calculates the optimum tour as 

 if the start-stop points were very close together 

 when, in fact, they are not. 



There is no requirement that distance be the 

 optimization parameter. Factors such as cost, 

 time, or suitable weighted combinations of other 

 variables could be used to compute a cruise track 

 considered optimum for specific user require- 

 ments. Also, there is no requirement that the prob- 

 lem be symmetric or Cartesian in nature. For 

 example, the distance (cost, time, etc.) in going 

 from station A to station B need not be equal to 

 that from station B to station A. Applications of 

 these characteristics and other distance matrix 

 manipulations include: 



1) The "cost" in going from station to station in 

 the presence of strong currents, such as the 

 Gulf Stream, could be adjusted. "Downstream" 

 directions from station to station would be 

 given preferential status for computing the op- 

 timum cruise track. 



2) In some situations, it may be desirable to group 

 selected stations to be sampled preferentially 

 as a subset or subsets of the total station pat- 

 tern. This might occur if certain sampling 

 areas had a higher priority than others because 

 of biological and/or environmental considera- 

 tions. 



3) Actual curvilinear distances between stations 

 could be entered into the distance matrix when 

 sampling in areas near the coast. This would be 



done for station pairs connected by a straight 

 line that passes across land. 

 4) If the number of stations exceeds the present 

 150 maximum allowable (300 with extended 

 core storage), and it is possible to divide them 

 into subgroups, the problem is limited only by 

 CPU restrictions. 



Many variations of the optimum cruise track 

 theme could be solved with this program and the 

 requirements are usually unique to a particular 

 problem or investigation.'* The examples demon- 

 strate the types of problems that could be solved. 

 Simple problems, such as those solved for the 

 groundfish survey, can be improved about 7% over 

 manually produced cruise tracks. 



Improvements obtained using the optimized 

 cruise track for the cited application are not 

 dramatic, but would be significant over a long time 

 period and/or extensive cruising distance. The 

 program eliminates selecting stations from ran- 

 dom number tables and hand plotting the cruise 

 track, which may require several man-days 



Literature Cited 



Oper. 



BELLMORE, M., AND G. L. NEMHAUSER. 



1968. The traveling salesman problem: A survey. 

 Res. 16:538-558. 



Lin, S., and B. W. Kernighan. 



1973. An effective heuristic algorithm for the traveling- 

 salesman problem. Oper. Res. 21:498-516. 



Thomas D. Leming 

 hillman j. holley 



Southeast Fisheries Center 

 National Fisheries Engineering Laboratory 

 National Marine Fisheries Service, NOAA 

 NSTL Station, MS 39529 



■^Inquiries regarding possible uses and applications of this 

 system should be directed to the Director, Southeast Fisheries 

 Center National Fisheries Engineering Laboratory, National 

 Space Technology Laboratories, NSTL Station, MS 39529. 



714 



