Method for optimizing resource allocation and an apparatus using the method

A method of optimizing allocation of plural usable physical resources to allocation objects, executed in a resource allocation planning support system using a resource allocation optimization model stored in a memory and processed by a calculating unit, said model represented by linear allocation constraints, an allocation state description of each allocation object represented by 0 or 1 variables assigned to all divided regions covering the entire allocation base region and an objective function described as a quadratic summation of resource stacks allocated to each of the divided regions of each said allocation object, comprising the steps of:

• inputting input data designating available physical resources, said linear allocation constraints, a specification of said allocation objects, and an initial state of said each allocation object;
• iteratively selecting, using a means for optimizing, a variable describing a resource allocation state of said resource allocation optimization model to be changed by a variable selection rule based on an evaluation vector representing present resource allocation stacks of all allocation objects, characterized by using an evaluation index for evaluating a value change of said objective function brought out by variable selection, defined by an inner product of an evaluation vector representing present resource allocation stacks of all allocation objects and a vector representing the new resource allocation corresponding to a newly selected variable as to an allocation object, and iteratively selecting variables to be renewed to minimize said evaluation index;
• renewing, using said means for optimizing, a variable of said resource allocation optimization model using the variable selected in said iteratively selecting step;
• judging, using said means for optimizing, termination of said iteratively selecting step for selecting said variable by evaluating whether improvement of said objective function is obtained by using said evaluation index, said objective function being made of said resource allocation stacks of all allocation objects;
• storing variables finally obtained when said objective function cannot be further improved, as an optimized resource allocation in said memory;
• and
• visually outputting a plan for allocating said physical resources to said each allocating objects by using said optimized resource allocation stored in said memory.

A resource allocation planning support system for optimizing allocation of plural usable physical resources to allocation objects using a resource allocation optimization model stored in a memory and processed by a calculating unit, comprising:

• a means for inputting input data designating available physical resources, linear allocation constraints, a specification of said allocation objects, and an initial state of each allocation object, said input data for controlling said optimizing allocation of plural usable physical resources;
• a memory storing said input data;
• a means for optimizing said resource allocation by iteratively selecting a variable describing the resource allocation state to be changed in a resource allocation optimization model represented by said linear allocation constraints, an allocation state description represented by 0 or 1 variables assigned to all divided regions covering the entire allocation base region and a quadrative objective function based on a variable selection rule based on an evaluation vector representing present resource allocation stacks of all allocation objects, characterized by using an evaluation index for evaluating a value change of said objective function brought out by variable selection, defined by an inner product of said evaluation vector representing present resource allocation stacks of all allocation objects and a vector representing a new resource allocation corresponding to a newly selected variable as to an allocation object, and including means for selecting variables to be renewed to minimize said evaluation index, for renewing a variable of said resource allocation optimization model corresponding to the variable iteratively selected by said means for optimizing, and for judging termination of iterative selection by said means for optimizing based on an evaluation index provided for evaluating whether an improvement of said quadratic objective function is obtained;
• a means for storing intermediate and final results of said optimizing process when said quadratic objective function cannot be further improved, as an optimized resource allocation in said memory;
• and
• an output means for outputting a plan for allocating said physical resources to said each allocating objects by using said optimized resource allocation stored in said memory.

A resource allocation planning support system for optimizing allocation of plural usable physical resources to allocation objects using a resource allocation optimization model stored in a memory and processed by a calculating unit, comprising:

• a means for inputting input data designating available physical resources, linear allocation constraints, a specification of said allocation objects, and an initial state of each allocation object, said input data for controlling said optimizing allocation of plural usable physical resources;
• a memory storing said input data;
• a means for optimizing said resource allocation by iteratively selecting a variable to be changed in a resource allocation optimization model represented by linear allocation constraints, an allocation state description represented by 0 or 1 variables assigned to all divided regions covering the entire allocation base region and a quadrative objective function based on a variable selection rule, characterized by using an evaluation index for evaluating a value change of said objective function brought out by said variable selection, defined by an inner product of an evaluation vector representing present resource allocation stacks of all allocation objects and a vector representing a new resource allocation corresponding to a newly selected variable as to an allocation object, and including means for selecting variables to be renewed to minimize said evaluation index, for renewing a variable of said resource allocation optimization model corresponding to the selected variables, and for judging termination of said optimization of said optimizing process based on an evaluation index provided for evaluating whether an improvement of said quadratic objective function is obtained;
• and
• a means for outputting intermediate and final results of said optimizing process as a plan for allocating said physical resources to each allocating object;
• wherein said means for inputting input data for controlling the optimizing process includes means for itemizing and visualizing said inputted information and for furnishing guidance in the inputting of said inputted information.